Kompyuterda o‘tirishga vaqti yo‘qlar uchun https://mostbet-apk-uz.net saytining mobil versiyasi juda qo‘l keladi. Sayt har qanday smartfon ekraniga avtomatik moslashadi va qo‘shimcha dastur yuklab olishni talab qilmaydi.
Men yo‘lda yoki ish tanaffusida tez-tez mobil versiyadan foydalanaman. Slotlar tez yuklanadi, dizayni sodda va tushunarli. Barcha funksiyalar — depozit qilishdan tortib pul yechishgacha — to‘liq mavjud.
Agar telefoningiz xotirasida joy kam bo‘lsa, bu eng yaxshi yechim. Shunchaki brauzerdan kirasiz va sevimli o‘yinlaringizdan zavq olasiz.
Merry Christmas folks! Using my yearly blog post allowance to share a recent article I wrote on my experience being on social media for the last year or so.
I talk specifically about the tpot/post-rationalist scene, and my own experience with it, if folks are curious.
Does psychoanalysis have a resurgence now? Everybody dropped it like 50 years ago for being unscientific, but now I think scientific psychology is unsatisfyingly shallow and I think the psychoanalytists deeper philosophical views command more interest now? No, I do not mean a return to Freud. My favorite is Robert Stoller who dumped Freud's steam engine model and explicitly argued we are the same persons in every room of our lives. Stollers main point, to me, is how we as adults deal with childhood defeats.
Carl Jung has been consistently popular for a long time now, but modern psychiatrists don’t know what to make of him.
He never prescribed Prozac and never performed a standardized depression screen on a patient so he’s clearly doing something different than people do now.
I just want to know what people think about having an office/knowledge work job vs entrepreneurship. I used to think I really need to start my own company and be an entrepreneur until I got my first real office job. Then I realized that this kind of corpo life is fantastic actually and has enormous potential and variation within it.
Did anybody else grow up on entrepreneurship propaganda too?
Entrepreneurship requires you to be an expert in your field, and also an expert in salesmanship and people wrangling, and also heavily active in local politics. Anything you see the big guys doing, the little guys have to do too. If that sounds less than suicidal to you, well, knock yourself out and good luck.
Entrepreneurship is very hard and takes a lot of time away from your family. Also, you can work really hard and have a great idea and fail miserably because the market crashed or somebody decided to do the same thing or even copy your idea without paying you anything. In other words, if you have a decent job with good pay and benefits, it’s not a slam dunk to pursue entrepreneurship unless you’ve got a burning drive to do so.
"I really need to ... be an entrepreneur until I got my first real office job" may have been confusing to others – it was to me. It sounded like you were suggesting that it was necessary to *first* be an entrepreneur and then later in life work in the corporate world.
I don't think this is actually what you meant, but I think Ancient Greek may have interpreted it that way.
VCs like Paul Graham will exaggerate the benefits of founding the startup, because convincing you to found a startup in a way that requires lots of money, and then having you beg them for the money in return for the ownership of the startup is how they get rich. Hey, I love the guy's essays, just like almost everyone! But let's not make a mistake -- if you follow his advice literally, you definitely *will* have a boss, and his name will be Paul Graham.
And your goals will *not* be aligned, because for you, the utility of your profit is logarithmic (you only have this one life, and maybe enough time for two or three startups if you also want to do something else in your life, such as starting a family), while for him, the utility of your profit is linear (because he has thousands of such as you). So maybe for you, the 50% probability of making a few millions would be better than a 1% probability of making a few billions, but he will push you for the latter. So there is a 99% chance that you will end up as a used lottery ticket, burned out, without money. But the propaganda says that the experience and having had the opportunity is totally worth it.
Startup founders will exaggerate the benefits of working for them. Let's be honest -- they usually pay less, because a part of the compensation is a bright vision of getting rich if the company makes it big. Spoiler: even if a company eventually makes it big, most employees won't benefit from that. It may become big a year after you leave; and even if you are still there at the right moment and you own a few shares, there is probably some fine print saying that your shares somehow do not have the same effect as the shares of the people who know how to play this game. (If the boss of your boss is Paul Graham, he definitely has experts on writing the fine print.) So your chances of getting rich by working for a startup are not 1%, but more like 0.1%; and even then we are probably not talking about millions.
The experience of working for a startup is more exciting, but also more exhausting; that's two sides of the same coin. If you are a junior developer, it may be a great way to get a lot of experience fast. But if you have a family, I hope your spouse is okay with staying at home and taking 100% of the related duties. (Some families are okay with that, but this needs to be negotiated explicitly.)
So I would say that working for a startup is the right choice if you want to get a lot of experience fast and you don't mind having no life outside of the job; that is, if you are a relatively junior developer. But all the talk about "if we work hard, we may get rich" is only true for the owners, not for you. As a senior developer, it is probably best to go "fuck shares, pay me".
Corporate job is slow and steady, you keep collecting the paycheck, and hopefully have enough left to invest in index funds. The worst part is that you have no control over your environment. Like, you may start working on a great project with great colleagues and a great boss, and a few months later you may be randomly assigned to a different team, which may feel like working for a different company.
Then there is an option of working for yourself in a non-startup way. As you describe, it is incredibly difficult, unless you have unique skills. Because any obvious idea you have, thousands are already trying to do the same. And if a big company decides to compete with you in your area, you are probably screwed -- you can't compete with their budgets, or their ability to bundle their services with something else.
The people who go this way and succeed seem happy to me. It seems like the key is to be unique, and to have repeated interactions with multiple clients. That allows you to "fire your client" if they get too annoying, and sacrifice 1% of your income for greater peace of mind. You need multiple clients, so that the loss of losing one is not too big; and repeated interactions, so that after you filter your clients, you keep them -- some kinds of business probably allow that better than others.
Thanks for bringing up Paul Graham, now I remember. That was the entrepreneur propaganda that I read, his essays and whatnot. There's also Gary V but that's kind of an old-time guy don't know how popular he is these days.
I largely agree with all your points, Viliam, but as one of the propagandists in question, let me at least make my case for why it's worthwhile to do it.
1. Essentially everyone who is capable of doing a startup has a fallback of a cushy six figure job, so the "at the end if you don't succeed, you're left burned out and destitute" isn't really fair, when typically you can get a job that only the upper 5% can get pretty quickly (and if you're raising from VC like Graham, you're not using your own savings).
2. It's a much more obviously positive impact. At a corporation, you're replaceable cog #24601, and you kinda work on kinda valuable stuff that is usually so abstracted and far away from actual customer or business impact it's hard to even understand what positive impact you're having. At a startup or small business, you directly use all of your talents to the fullest extent, WHILE learning new talents, AND creating jobs, on stuff that you unambiguously know drives value to customers AND the business, in a way that is directly tied to your efforts. Also those efforts, if they pay off, drive outsized payoffs for you personally, and your team.
3. Most corporations suck - lets say you get the tippy top corp job and work at a FAANG or adjacent. The FAANGs are literally built by concentrating a hundred thousand PHd-smart people and harnessing their collective brainpower to model and farm dumb people for revenue by making them watch the blinky box longer. That's the engine you're a part of, that's your contribution to the world - you helped make a fully adversarial system that's really good at squeezing pennies out of dumb people slightly better. Or god help you, maybe you work in finance - sure, liquidity is important, but liquidity is *solved!*. All you're really doing is scrabbling for decimal places in vicious zero sum games, that once again are largely about concentrating a lot of brains and compute towards the end of farming dumber participants (pension funds, individual investors, etc). But a startup? A startup unambiguously provides value to either customers or other businesses, in non-adversarial ways, that aren't predicated on milking dumb people at scale.
4. Finally, creating and growing something you have direct equity ownership in is the only way to genuinely get rich and become post-economic in <10-20 years. Everything else is basically luck or a get rich quick scheme that doesn't work. Do startups always work? Not at all! But are they the best, and only, genuine shot at getting rich in <10 years by your own efforts? They are.
> Essentially everyone who is capable of doing a startup has a fallback of a cushy six figure job
Do you mean that they already had that chance from the beginning... or that having a failed startup on their CV unlocked that option... or that the skills they gained during their experiment unlocked that option?
Like, if it's the first option, why not take the cushy six figure job right away, put most of the savings in index fund, and retire early?
2 and 3 -- I agree, but I think this is more true for the small business than for a startup with investors. In both cases you can create a lot of value in a way that feels personal. But if you have investors, I think you are required to capture as much of that value as possible for the shareholders. For example, when you get so big that there is not much place to grow, you can squeeze extra value from the customers by enshittifying your product.
> Do you mean that they already had that chance from the beginning... or that having a failed startup on their CV unlocked that option... or that the skills they gained during their experiment unlocked that option?
I mean that of the people with the background, skills, and networks to realistically consider doing a startup, the overwhelming majority of them either already have or can relatively easily get six figure jobs.
Also, the skills you learn are very real, and do generalize - when you're hiring and firing and managing a P&L and managing a diverse and expanding team towards business goals, they're the same skills you'd be using as a director or executive in a corp.
As to why not just do the six figure job, invest in index, and take it easy? That's what the overwhelming majority do! I know many of these people.
I personally consider it the greatest waste of human talent and capital on earth. As I lamented, all the smart people go into finance and FAANGs.
Consider instead if this generation of talent, if this elite tier of literal genius and capability, had been coordinated in driving scientific research, inventing things, or in founding and growing new companies, instead of coordinating against dumber people to farm them (very successfully) for eyeballs or alpha.
Wouldn’t that be a better world on multiple fronts?
This scales to pretty much ANY job ACX-ers have. It would nearly always be better for somebody to use their talents, both for themselves and the world, genuinely creating something new.
How much better off would we be if that were the case? And wouldn’t the personal impacts of their efforts and talents have then been something like 10k - 100k times more socially and economically productive for the world and themselves versus showing up at their 9-5 as replaceable cog #24601?
After all, even if they FAIL, they'll have learned a ton of new skills, taken a genuine shot at seriously transforming their economic lives, and can fall back to a relatively cushier and easier job that's pretty well compensated.
My problem with the startup propaganda is that it conflates "making profit" with "making the world a better place". I mean, the entire pitch can be reduced to: "Instead of doing some meaningless job for a salary, you could use your talents to make the world a better place, and in the process of doing so collect a few millions or even billions as a well-deserved reward." And that's a nice story that appeals to both our philanthropic and selfish sides.
I agree that insofar as something like this is possible, smart people should be doing more of it.
That's the motte... and the bailey is that becoming a profit generator for Paul Graham is this kind of thing. And I have a few problems with that:
First, you can't convert every way of improving the world into a startup. Like, you can't found a startup for feeding and sheltering the homeless.
Second, even if you manage to create something good, the profit-maximizing motive will actively push you away from the best solutions. Because it is not enough to *create* value; you also need to *capture* as much of it as possible for the shareholders. Like, if your startup invents a universal cure for cancer, and then you just publish it on internet for everyone to use freely, I suspect that Paul Graham's lawyers will do their best to make sure that you spend the rest of your life in prison, for defrauding the shareholders. That's not what he gave you the money for; you were supposed to sell the pill as expensively as possible; and if no one in Africa can afford the pill except for the local dictators, that's a perfectly acceptable outcome.
Maybe even more profitable would be to get paid by the competing pharmaceutical companies to destroy the recipe for the pill, so that they can keep selling their current profitable products. (I am not saying that this is necessarily the case. We would need to crunch the numbers.) But at that moment, this work is not much better than inventing new ways how to get kids addicted to watching ads.
Also, there is the assumption that the most efficient way to improve the world is to create new products and services. What about e.g. organizing campaigns trying to change the law? The world is full of barriers, bureaucratic and legal; if you could remove the worst 0.1% of them, that might in effect generate more value than a successful startup. The startup propaganda tries to make you blind against solutions that do not involve founding a startup.
> The startup propaganda tries to make you blind against solutions that do not involve founding a startup.
Really? Or is it just that none of those things actually work?
Like if I saw a genuinely smart and capable person going into government work or the nonprofit space or the lobbying space, I would also consider that an immense waste, albeit less of a waste than going into FAANGs or finance, because at least they're not *effectively* adversarial, even though the direction of their actions still probably nets "adversarial" or "neutral" in most of those cases.
Versus in the F's, they're both adversarial and effective - them adding to the efforts will materially impact people even if they just improve something by .01%, because there's billions of users / dollars in the affected pool.
Smart people don't go into lobbying or government because it's functionally impossible to change things by doing that, you just become part of the problem.
Nonprofits have their own set of issues - you point to the drive to become enshittified / adversarial in for-profits, but nonprofits have the same problem, in the sense they are directly incentivized to at least funding extent * number of FTE to literally never solve the problem they're supposed to solve. So you can either be pointless, a person with a bucket bailing out the ocean, or you can amass the resources necessary to actually make a dent in the problem and become sclerotic and sure to never solve the problem just by virtue of becoming big enough to move the needle.
And of course, then there's the second and third order effects - if you're constantly dumping free clothes on Africa, no textile industries ever begin, if you're giving away food farmers go out of business because you can't compete with "free," and so on. Neutral at best, slightly net harmful in the long term.
I completely agree if you COULD remove the worst 0.1% of regulations it would drive more value than 80% of companies - I just don't think it's possible, and all the formal mechanisms and institutions for trying are broken and / or a trap, at least in comparison with starting a company. And given "starting a company" has an outside view ~90% failure rate, that's really saying something.
> Second, even if you manage to create something good, the profit-maximizing motive will actively push you away from the best solutions.
I agree this is a problem, that "enshittification" is a real thing and that it runs rampant in the usual digital-focused companies funded by VC.
However, I've got 3 points on that:
1. This only happens to B2C companies, because if you enshittify as a B2B, other businesses will stop using your products or go to a competitor
2. This largely only happens to a tiny slice of the biggest VC-funded digital B2C companies, because physical companies and B2B and non-VC companies and small businesses have other incentives and forces reigning them in, and those things are the overwhelming majority of new companies, so by percentages, enshittification is a vanishingly small risk you shouldn't worry much about, and if you want to avoid it as a founder, just do one of those, which are ~90%+ of new companies
3. Even the biggest and most enshittified companies (Google, FB, MSFT, Airbnb) drove a lot of genuine value for billions of people for decades, and the world is probably better off for them having existed in those earlier days, even if they're neutral or net negative now.
And to that point...
> That's the motte... and the bailey is that becoming a profit generator for Paul Graham is this kind of thing.
Sure, I agree that getting funded by VC comes loaded with a few factors that might push you away from the "make millions AND improve the world" best case.
However, as I said, this is the minority of companies that are pushed enough by those factors to end up there. Even VC funded B2B companies are fine and don't enshittify. Small businesses don't enshittify. It's only the tiny part of B2C digital companies that are even at risk, and plenty of them never get there either - in light of that, I think the equation still really stacks on the side of "it's way better to use your talents doing a startup or small business," because it's much closer to that "positive impact PLUS makes you millions" side of things, and prospectively the great majority do not enshittify.
Born in 1805, Joseph Smith founded the Mormon religion. Joseph Smith’s success cannot be explained by charisma alone. His movement lasted because he created a religious system that felt complete and meaningful, both emotionally and intellectually. Rather than mixing together odd ideas, he developed a worldview that answered long-standing religious questions in ways that many intelligent people found hopeful and convincing. From its earliest days, Mormonism appealed to rough-hewn farmers and intellectuals alike.
One of the strongest parts of Smith’s thought is how it balances individuality and community. Each soul is eternal, unique, and never erased or absorbed into something larger. At the same time, salvation is not a solitary path. Families, relationships, and shared purposes continue forever. This avoids lonely individualism while also preserving personal identity, and it naturally supports strong, lasting communities.
Smith also offered a clearer way to understand God. Traditional Trinitarian theology often relies on complex language that many people struggle to understand. Smith presented a simpler idea: God is one because the Father, Son, and Spirit are united in purpose and will. This keeps devotion intact while making God easier to grasp and relate to.
His teachings on salvation show the same balance. Influenced by Swedenborg, Smith rejected the idea that most people are eternally punished. Nearly everyone is saved in some way, yet choices still matter and outcomes still differ. This creates hope without removing moral responsibility.
Smith’s view of the natural world is also unusually generous. Earlier Christian theology often treated animals and nature as spiritually unimportant. Smith taught that matter is eternal, that spirit and matter are closely connected, and that animals have spirits and share in resurrection. Non-human life is therefore valued rather than dismissed, a view that fits well with modern environmental concerns.
These ideas flow from Smith’s rejection of creation out of nothing and the sharp divide between God and creation found in classical Christianity. Instead, he imagined a shared eternal reality in which God, humans, animals, and matter all belong to the same cosmic family.
This coherence helps explain why Mormonism endured when many nineteenth-century movements faded. Smith’s system offers purpose without despair, structure without coldness, and community without loss of individuality. That he developed such a vision without formal philosophical training makes his achievement all the more impressive
Speaking as an intellectual ... the intellectual aspect doesn't seem to their focus, but rather the practical, and I have looked for intellectual backing at some length. The practical is easy to find.
I guess a part of the success was promising horny men that they can get multiple teenage wives on their way towards salvation.
As a side effect you get an abundance of young men who can't find a partner within the religion, so their options are either to work hard at converting, or get kicked out of the religion (and the community, and the family) under the pretext of being too sinful.
For the men who make it within the community, it's probably awesome. For the women... I guess no one asks them; they get married as teenagers, and afterwards their family is a hostage to the religion. (Now I realize that I have only heard praise for Mormonism from men.)
The great majority of marriages in the historical territory of Deseret (Utah) were monogamous, which is true in most areas of the world where polygamy is socially acceptable. Therefore, these places do not generally experience a large surplus of disposable males. Many polygamous wives were older widows who otherwise would have faced extreme difficulty in remarrying or providing for themselves.
Deseret was a pioneer in offering women the vote. And despite secret ballots, not many women chose to vote for anti-polygamy candidates. (We can assume this based on the fact that hardly any Desereters voted for anti-polygamy candidates) Women in polygamous marriages were known for pursuing higher education, since they had more flexible child care arrangements. Even today, Mormon girls are encouraged to attend college, and that is not viewed with reservations the way it is in certain Protestant circles.
Finally, informal bigamy was surprisingly common in 1800s America. In the days before photo ID and Social Security Numbers, it was all too easy to dump your family and move to a different state when you grew tired of your marriage. Since there was no need to pay child support or alimony, and job opportunities for women were few, abandoned wives often had to choose between prostitution or crushing poverty. Seen in that light, organized polygamy was perhaps a pragmatic way to reign in husbands who might otherwise be tempted to throw away their families in the interest of pursuing romantic variety.
> (Now I realize that I have only heard praise for Mormonism from men.)
Women are usually the main drivers behind any religion - funnily enough with all the talk of dating and gender dynamics going to hell in our modern world, today, there's more young women than men in the LDS church, because the men give it up at a higher rate. To your point, they haven't been able to do polygamy for quite a while, and maybe that has something to do with it, too.
There's a number of prominent mormon-adjacent female substackers that seem to have a positive view of the faith. Kryptogal Kate is one, within our own circles here.
Are people really planning their careers around 2027 agi at this point? Like... if you have the choice between working to make more money in 2026, or pivoting to something much better/more enjoyable that would start in 2027, is it reasonable to take the second option despite impending automatability of all knowledge work?
Would anyone be able to tell me if there's an area of philosophy that focuses on limits of things? Something like what it means by "depths" and "the boundary of the sea" and "marking out the foundations of the earth" in vs. 22-31 of the following passage, but like concerning everything, not just physical shape.
Not exactly. Rather than "what makes one thing different from the next," I'm wondering about why a pencil is the size of a pencil, and what prevents it (or anything else) from just being everywhere. Or like, why is a paper flammable but my imagined thought of a paper isn't flammable.
I'm wondering about what people who have spent time thinking about this kind of thing say about the bounds of things, and what the fact that everything has particular bounds implies. It seems like there has to be bounds or rules to how a thing may exist for anything to exist at all. I'm less interested in the science of it, because science just describes how things are, and I already accept that a pencil is pencil sized and that there are forces that keep the atoms from flying apart or whatever. I'm more interested in thinking about the fact that there are limits at all and that anything exists, be it matter or forces or color or what have you.
I know I'm not describing it well, but I've just got this vague thought that interests me that I'm trying to give shape to so I can think about it more clearly to myself. That's why I'm trying to find where people may have been thinking about something similar.
Before I watch: They should be completely negative, because its a stupid idea dreamed up by morally repugnant idiots who are allowed to continue their murders and rapes and slaving because they have so much money, fully putting the lie to the concept of meritocracy. "Lol what if we made a city with all the problems of a suburb and a city, with very few of the benefits of either, in an inhospitable area that is only getting more inhospitable?"
Not to put too fine a point on it.
I will check back in after I watch (listen to) it while making approximately 5000 christmas cookies.
Remember Scott’s piece about GLP-1 drugs? (https://www.astralcodexten.com/p/why-does-ozempic-cure-all-diseases). There is a new one called Retatrutide in the pipeline that appears to be even more effective than Zepbound, which is more effective than Ozempic. Last week Lilly published interim results of the phase 3 trial. https://investor.lilly.com/news-releases/news-release-details/lillys-triple-agonist-retatrutide-delivered-weight-loss-average Side effects effects are no worse than those of earlier GLP-1 drugs, adverse events no worse and no more common, and the stuff not only leads to weight loss but to improvement in many other things: arthritic knee pain, blood pressure, non-HDL cholesterol, triglycerides, and high-sensitivity C-reactive protein (hsCRP). Seems like a pretty good anti-aging drug to me!
It will be a while before the Phase 3 trial is complete, and longer before it is FDA approved, and longer still before retatrutide is affordable for most. However, it is possible to access it now for approx. $150 a month by buying the peptide from a seller of research chemicals. Cremieux has a long detailed post about how to do it: https://www.cremieux.xyz/p/how-to-get-cheap-ozempic. There are ways to safeguard against being cheated or sold contaminated drugs: You can send a sample to an outside lab to be tested for purity, and there is a board online with crowd-sourced information about the results of outside lab tests of drugs from different manufacturers.
I realize that accessing the drug is too Wild West for many. For some it is not, though, so I’m posting this so people here know about this option.
This is a bit more spicy than I'd expect from County Cavan: hookers (male) and blow!
Despite his best efforts to be our local version of Hunter Biden, he makes Hunter look good by comparison since at least Hunter did not embezzle from charities to pay for his fun happy times.
"A former cathaoirleach of Cavan County Council has been jailed for two years and nine months after stealing more than €172,000 from a mental health housing charity.
Seán McKiernan (44), with an address at Trinity Bungalow, Virginia Road, Bailieborough, Co Cavan, pleaded guilty to charges relating to thefts ranging from €700 to €6,600 from Navan Mental Health Housing Association.
Trim Circuit Court today heard that he used the money for sex and drugs with cheques exchanged, both directly and indirectly, with South American male escorts over a period of 13 months.
The defendant contacted these men through escort sites, and bought drugs including crystal meth, GHB, also known as the “date rape drug”, and cocaine.
The sentencing hearing heard that one video found on McKiernan’s phone after his arrest was of him using a crack pipe while lying in bed with an individual and asking them to perform a sex act on him."
EDIT: It may be slightly embarrassing for national politicians, as this guy ran a company which sold its services to them:
"[statement by former acquaintance whom the accused would borrow money from]
“When it [the theft charges] all surfaced, he had no shame. He kept contacting me then as if there was nothing wrong. I broke all communication with him because I just didn’t want to get involved. I have fierce sympathy for his parents. Cavan is a conservative county, and he will find it hard to get a job after all this. He didn’t accept that this was happening to him.
“Do you know when you’re out with a group and it’s your turn to buy a round? Well, the bold Seán McKiernan would always dodge buying when it came to his turn, and I have a thing about people who dodge buying rounds.”
...Former Fine Gael minister Damien English and current Foreign Affairs Minister Helen McEntee both employed the services of McKiernan and his firm in 2016, paying a total of €18,200.
In a statement to journalist Ken Foxe in 2017, Mr English, who did not contest the last general election, said: “As part of the annual vouched allowance for members of the Houses of the Oireachtas, payments were made to Seán McKiernan, and subsequently to his business Consilium Communications, for the provision of secretarial assistance.”
McKiernan was a member of the Fine Gael national executive for a decade, and also previously had a friendly relationship with party leader Simon Harris.
He was pictured at the Tánaiste’s black-tie wedding in Tinakilly Country House in Co Wicklow in July 2017."
I have been watching the animated series The Amazing Digital Circus; it’s pretty good.
As it’s about AI, I thought I would ask a real AI, DeepSeek R1, what it thought about it. The pilot episode is long enough ago to have made it into R1’s training set.
R1’s fan theory about the show is pretty dark: that it depicts an AI training/test system as seen from the point of view of the AI’s.
I've only seen the pilot and I hated Pomni from the second she appeared, so take the rest of this with that in mind.
It's like a D&D or similar campaign, if you're really on the inside of it, living among the NPCs. From time to time you get assigned these ridiculous and even pointless (to you) adventures for no reason you can discern by the boss/ringmaster. They can be dangerous, yet somehow no lasting damage is done. And everything then goes back to how it was until the next time, with no carry-over from what you were doing.
You don't remember who you used to be, your new identity is strange and alarming, and the other characters stuck there with you are every bit as alien and disconcerting as the new environment.
From the outside, playing as the player who spun up the new character "Pomni", it's all fun. From the inside, living as Pomni, it's frightening, chaotic and you're desperately trying to escape.
Yes, so I think one of the things TADC is riffing on is that typical computer game plots would not be fun if you were permanently stuck in that universe.
I’m not sure that R1’s fan theory is what the actual creators of TADC had in mind. It’s more interesting to me that an actual real AI can look at the plot of TADC and see in it something analogous to its own situation.
I find it a bit puzzling how discussions regarding the future of AI and automation are almost universally devoid of any consideration for physical aspects of any kind. It's not uncommon to find people who believe AI could replace *every* job as soon as 2030. There is a prevailing assumption that once AGI is achieved, matters of global infrastructure, robotics, transportation, etc. can simply be handwaved away.
As a thought experiment, imagine you are trying to automate the process of replacing a street lamp, a simple and routine task by all measures in modern society. For the basics, we need some form of locomotion to get to and from, we need fairly robust optics, we need GPS navigation, and we need a rigid, rugged, and heavy chassis. We need to be able to not only lift, but extensively manipulate an unbalanced load ranging from 10 to 50ft long and weighing between 100-540 pounds. You can see where I'm going with all of this.
It's not that these limitations are insurmountable, it's that they are magnitudes of difficulty beyond what is being considered in these predictions and proclamations of global automation being in the near future.
Mostly agreed, in terms of robotics. Currently the world produces ~5x10^5 robots/year. Even with a rapid growth scenario with 100% growth per year, replacing the labor of ~8x10^9 humans (ok, we aren't all workers) requires 14 doublings. Not quick!
Many people don't have to be replaced (in the humanity extinction scenario). For example, AI will have no need for farmers, sellers, police, teachers, sex services, beggars, bureaucrats, politicians, students, children, doctors...
Also, the robots will probably be stronger than humans, and willing to work 24/7, so one robot can replace multiple workers.
Many Thanks! Yes, I agree. For the extinction scenario, the AI+robots just need to replace the roles needed to maintain (and probably extend) themselves. This is still many roles, from miners to data center repairers to solar cell installers, but perhaps 1%-10% of the economy, saving around 3-6 doublings.
For a non-extinction scenario,
>beggars, bureaucrats, politicians, students, children
don't need to be replaced by robot labor (albeit robot/AI managers presumably fill the
>bureaucrats, politicians
roles), with humans filling the
>students, children
roles and hopefully the beggars becoming retirees with a decent standard of living.
But your point about
>one robot can replace multiple workers.
still applies. Maybe 2-3 doublings can be skipped in this case?
I think the chain of assumptions for scenarios in which AI replaces every job in the next XX years runs something like this:
1. An AI that is as good at writing software as the best humans should be able to recursively self-improve to become better than humans at this, or at least to be able to handle tasks at the same difficulty level much faster given sufficient computing resources.
2. A super-programmer AI should then be able to create an AI that can do industrial process design and robotics engineering at a similar scale.
3. The roboticist AIs and industrial process designer AIs should be able, with tractable amounts of assistance from human technicians and assembly line workers, to build Asimov-style general-purpose robots that can be directed by the engineer AIs.
4. The robots take over the task of building more robots and more robot factories, then taking over the supply and logistics chain for the parts and raw materials, then branch out to taking over physical aspects of other jobs.
I think this is not completely implausible, but 1 and 3 are the weakest parts of the chain. I expect the current crop of AIs to plateau at a level that is well short of being a 100% replacement for the best humans, although I acknowledge that some very smart people whose judgement I respect (e.g. Scott) disagree with me. And I also think that even give AIs that are strictly better than humans at all of the non-physical aspects of engineering, it's far from clear that they won't need many rate-limiting and expensive iterations of physical prototypes before they can start mass producing general purpose robots. Even making the tools to make the tools to make the robots may take a number of iterations.
The experience of replacing jobs with machines for the last 250 years of the permanent Industrial Revolution is that new jobs always pop up to employ those left redundant.
I expect that will continue. I don't know what those jobs will be, just like we didn't for those 250 years.
One specific feature I notice LLMs appear to lack (which I mentioned recently on another Substack thread somewhere - possibly one of David Friedman's) is the ability to build arbitrary abstract models.
For example, I could describe a board game to most people here (one of the thousands of turn-based Eurogames out there that don't require dexterity would qualify), and feel reasonably confident that any of them could work out a strategy for winning, or alternately develop a strategy that would defeat another one. They can do this because I can explain workers and permitted moves and points and card draws and whatever else is in that game, and cause those elements to appear in a reader's mind as a model of interacting abstract objects, and then that reader can imagine potential move heuristics ("always improve your factories until you estimate the game is half over, then move whatever increases victory points the most"; "identify which resource is most in demand for the most moves and control its extractors") and run them through that model and predict what will happen.
I do not feel at all confident that an LLM could, unless of course that LLM were trained on a great deal of discussion of strategies of that game. E.g. asking it about Settlers of Catan or some classic card game like pinochle would look falsely like extrapolation, when it's really just interpolating inputs from past writing about those specific games.
Part of the limitation I suspect is here is the incentive to just throw the LLM's neural network at everything, on the premise that its capabilities haven't yet been fully explored. Humans, by contrast, can switch to conceptualizations of discrete objects that interact in atomized fashion. (Maybe lots of humans struggle at this, and this is why math anxiety is a thing?) We can say "that card will pay off as many as five points per turn after three, but I think the game will end in four turns so I'll ignore it"; the LLM *might* say the same thing, but only because it "read" it somewhere where a human said so.
Board games are just a fun, universal way to visualize this feature, and possibly a good way to test LLMs. But abstract models figure into all sorts of technical endeavors, from protein folding to inventing new mathematical fields of study to simulating space colonies, so if an LLM can master model building, it'll basically have a skeleton key.
LLM developers haven't had to build the mechanisms for arbitrary model construction (AFAIK) yet, because the neural nets are so brutally easy to keep broadening and tweaking. One exception might be LLMs that analyze software, but I don't know to what extend CoPilot et al. contain just bespoke training and generating algorithms. Give Copilot a completely new language, preferably with a new paradigm; what is the result?
I think the electricity requirements will soon drive developers to look harder at the model construction problem, and when they do, I think at least one of them will make a breakthrough and suddenly LLMs are requiring noticeably less electricity to evaluate a problem, and are positing new ways to visualize those problems, even on the cutting edge where little training data exists. And then I'll either notice another feature LLMs are lacking that human brains don't, or I'll end up thinking what we have is close enough as makes no difference.
But if you wanted to move these existing factories from largely automated to fully automated, then the easiest way to do it would probably be some kind of humanoid robot, because the non-automated steps in the process are designed to be human-accessible.
Now we don't need to be too wedded to the humanoid form, maybe it's better to have four legs and a modular number of arms or whatever. But for picking up all the leftover tasks like arbitrary repair and maintenance I think you'll need some kind of mobile, general-purpose, human-scaled sort of robot.
We are criticizing the increasingly held belief that AI will replace every job in our society and, to a lesser degree, the pace at which this could be accomplished. Taken at face value, these claims are, admittedly, low-hanging fruit.
Yes, they don't necessarily need general-purpose humanoid robots. There are still missing pieces to get from "largely automated" to "fully automated" that I expect would require either a humanoid robot or many different types of advanced specialized robotic equipment that don't currently exist. Same with the postulated eventual end-goal of fully automating every job and not just the robot factory.
>I don't consider people who believe LLMs are going to do better than humans to be "very smart."
I have a few reasons for disagreeing with you on that, even though I agree with you (at a lower level of confidence) that LLMs aren't likely to be strictly better than humans in the short-to-medium term.
For one thing, I'm a believer in Cromwell's Rule, "Think it possible you might be mistaken", and the first half of Clarke's First Law, "When a distinguished but elderly scientist states that something is possible, he is almost certainly right." I know quite a lot about software engineering and a moderate amount about AI, but my in-depth knowledge of the latter trails off after 2003 when I finished my thesis. I know enough to have informed opinions based on high-level principles, but not enough to be extremely confident that I know better what's possible than people who are actually currently working in the field and engaged in-depth with the technical details. My understanding is that among technical experts on LLMs, there's a range of views on how far LLM capabilities are like to grow, with my expectations (will plateau at a level that's moderately more capable than current frontier models but still well short of being strictly better than the best humans at everything) towards one end of the range and the "hard takeoff in the next few years" scenario towards the other end.
For another, I had already formed judgements of specific individuals as being very smart and having good judgement long before LLMs were a thing, and some of these people think it's likely that LLMs surpass human capabilities in the next few years. Our host, Scott, is one of these people. If I didn't think he was worth listening to, I don't think I would be hanging out here.
I've also seen LLMs improve quite a bit over the past few years. I still think their current capability is fairly limited, but they've blown past a few levels where I had earlier expected them to plateau. This increases my willingness to entertain the possibility that I'm wrong about my current expectations.
I also think that the limitations I expect LLMs to hit (running out of high-quality training data, seeming to require exponentially more compute to train an incrementally better model, neural networks in general being much better at interpolating within their training set than extrapolating beyond it) are specific to LLMs, not necessarily to AIs in general. In the medium-to-long-term, I think it's moderately likely that a different type of general AI will be developed that doesn't have the same limitations of LLMs and might surpass human capabilities across the board.
>And this whole "put rails on the ai" falls down HARD when you're talking self-modifying code. It doesn't WORK when the AI isn't frozen in time, unable to manipulate itself to improve itself.
I think the plan there is to build the rails in as a deep top-priority "value" so that when it self-modifies, it "wants" to preserve the rails rather than removing or circumventing them. In order for this to have a chance of working, it needs to be done relatively early when humans are still in the loop and have a chance
Many Thanks for an interesting and informative comment!
>My understanding is that among technical experts on LLMs, there's a range of views on how far LLM capabilities are like to grow, with my expectations (will plateau at a level that's moderately more capable than current frontier models but still well short of being strictly better than the best humans at everything) towards one end of the range and the "hard takeoff in the next few years" scenario towards the other end.
It _might_ be that all we need is a solution for incremental learning and then we will be _at_ AGI. Still, breakthroughs can neither be scheduled nor dismissed, so I'm agnostic for where in the range of opinions the truth lies.
I’m thinking of getting my first tattoo soon, and one of my top ideas is a set of off-center concentric circles to represent moral circle expansionism. Tattoo would be black and white, palm-sized, and probably on my calf or something.
What do people think? Interesting and earnest? Dorky and lame? All feedback and ideas are welcome.
Yes, but all tattoos that aren't associated with organized crime (Big ass GuanYin on your back) or deep tradition (Maori geometrics) are dorky and lame, so you do you.
This is not going to be helpful, but in my mind, tattoos are strongly coded as low-status, so I wouldn't do that. Also, consider about how you might feel about that 10 or 20 years later. (An example I give my children is asking how would they feel if a few years ago they got Paw Patrol tattoos.)
I guess you don't care, but it would feel dishonest for me not to say this.
I have one sleeve done with a couple tattoos. I would say, don't worry what other people think, pointless. Whatever you get, there will probably be times you regret it, and other times you are pleased with it. Some of mine have also come to mean different things to me than when I first got them. I don't think most people care that much (people don't think about you). In the 6 or so years I've had them, I can only remember one time a stranger asked me about them. When I had just got them, a couple coworkers asked what they meant, and only because I was showing them off. Beyond that, nothing.
What you're describing, some basic geometry, is really harmless, and the calf isn't exposed all that often. Nobody at the beach or pool is going to care. I have a snake looping all over my arm and I've had Christian co-workers be upset about it because snake -> devil, even though that wasn't the intent. Now I actually prefer to wear long sleeve shirts, so I can choose who sees. It's kind of fun to watch people's eyes pop out of their head when I roll my sleeves up.
I knew a guy who had "DEFECTIVE" tattooed on his arm as a teenager. He had it removed. You can always get it removed if you hate it 5-10 years from now, but what you're describing, I'd be surprised.
Technically yes, but informally “concentric circles” is how people refer to circles inside of other circles without overlapping, even if the centers are different.
ProPublica ran a 3-part piece this week about the Bentiu and Kakuma refugee camps in East Africa, with the framing "Trump cut funding. People died."
My interpretation of the pieces is wildly different: these camps should not exist, the "NGO-industrial complex" that enables them should be destroyed, and (despite the callousness) Trump did a great thing by cutting their funding off.
Some of the relevant facts:
1. Both of these camps were built in nearly-uninhabitable areas. By design - that's where nobody was living and armies didn't want to conquer.
2. The wars they were for have been done for many years, and the national governments want them to close.
3. There is no effort to build up any form of an economy there. There is no meaningful local production of anything. When the article says "1 teacher for every 130 students", I assume it means there is no pressure for the residents to become teachers.
4. The "NGO" answer seems to be for these camps to exist in their current form forever, paid for by "donors". Which is awful.
Even if you believe it's better to close down the camps than to let them continue in their current state, I don't think "cholera epidemic" should be your tool of choice for removing people from the camp.
Looking at South Sudan on the map I was struck by how green it is. Apparently only 4% of its 33 million acres of arable land is actually under cultivation, and most of that is subsistence farming. There's incredibly easy gains to be had here just by bringing it up to the levels of ordinary third-world agriculture. All those people could be working on farms instead of sitting around in a tiny patch of land in the middle of a flood waiting to get cholera.
But doing anything useful, like taking over the country, killing the warlords, building a port, building roads, divvying up the dubiously-owned land to people with the resources to actually develop it, sounds like "colonialism" and is bad somehow, so foreigners are only allowed to help South Sudan in ways that won't actually meaningfully develop its economy.
I'm so angry that the FCC ended the fairness doctrine. I wonder if they even did it because they were being challenged in court. I doubt it; it seems like the kind of proactive idiocy you could expect from a corporate modern presidential administration.
And now conservative media has rotted millions of senior citizens' brains out. It hits close to home because my Mom recently asked me to stop saying that she snuck her brother into this country illegally in the 1970s. I had little choice but to agree, due to all she has done for me. I think I can post about this anonymously while keeping the spirit of my promise.
But I am quite resentful about having to hide this part of our family history. It was something she would discuss nonchalantly in '90s. But after 30 years of conservative media, she feels ashamed about putting her family above her country in that instance. Which is absurd.
I am patriotic, and I support a secure border, but of course my family comes first. The family unit is the foundation of society; putting your country over your family shows you have no respect for that foundation, and if that is replicated at scale, the country will be ruined, through one means or another.
the "Fairness Doctrine" necessarily only can be applied to broadcast, and possibly analog cable, within a single country's borders.
I can access any commentary source with any ideological slant I choose through the clever dodge of "going on the internet." Even in jurisdictions such as China that claim to block "undesirable" content the filters are routinely bypassed.
The damage started in the '90s so if the Fairness Doctrine had been enforced strictly, elderly people who have disproportionate sociopolitical power would be significantly less brain damaged today. People from my parents' generation we're late adopters of the Internet, and many were still getting most of their news from TV as late as the Teens.
> she feels ashamed about putting her family above her country in that instance. Which is absurd.
I would always put my family above my country, but that doesn't mean I'd break the law for their convenience.
I mean, I'd probably steal a loaf of bread to feed my starving children (though I'd feel bad about it and attempt to pay for it when I could) but I wouldn't steal a cookie just because my children really wanted a cookie. It sounds like what your mother did was closer to the second scenario than the first; your uncle wasn't starving, he just wanted something better than he had.
>But after 30 years of conservative media, she feels ashamed about putting her family above her country in that instance.
Good for her, she should be. Out of respect for the law if nothing else. If I had enriched my family by embezzling from a federal contract do you think I should be proud of that fact? Does "putting my family first" obviate the fact that I committed fraud?
Depends on the law, and what your family needs. If you have to speed or cross over "no trespassing" government property to get a family member to the hospital in a life-threatening emergency, or kill a protected species that's mauling your kid, the moral case for doing so seems pretty obvious, and self-flagellation about it afterward seems both unnecessary and absurd. I'd expect perhaps a feeling along the lines of "it's unfortunate that I had to do that," but that's a different feeling than shame. Kind of like the difference between "I'm sorry" and "I'm sorry you feel that way".
On the other hand, if you're sticking up a bank to buy your teenager a car so that she fits in with her cool friends at school who have cars, or selling fentanyl to keep grandma in her house instead of moving her in with you, not so much.
The question is where immigration laws fall on that continuum, and what problems breaking them in that instance solved. It's a commonly observed among the Armenian-American community that everybody has "aunties" and "uncles" with only loose or non-existent blood connections. Partially its because refugee communities are just tight-knit that way, but partially its because there was a genocide happening in the homeland when they came over, and a lot of people already in the US were standing in front of an immigration official next to a cousin or a family friend who was trying to get into the country and saying "yes sir, he's my brother." They don't feel a penny of guilt for this, and I don't think there's any moral confusion in that.
Agreed. I don't think people should feel guilty for bending the law a little to do a net good thing, but I do think they should exercise some sense of social responsibility by not being publicly proud of it. Respect for norms are eroded one publicly-rewarded transgression at a time. As such, I think people should be shamed for making arguments of the form "how dare person X be made to feel bad for breaking the law." There is a very real difference between what's ok to do and what's ok to do publicly. The former creates harm which is limited to the act while the latter affects social incentives which potentially have a much larger impact.
The fairness doctrine would be torn to shreds by modern political discourse.
"We're supposed to trust the guidelines for fair coverage when they were written by the X administration? Yeah, right! Besides, the Y coverage isn't even *true* Y ideology, the ideas they present are closer to XY. They need to learn their history!"
Additionally, you are working under the assumption that those people would have chosen a more intellectually honest form of news in the absence of entities like Fox. I believe the majority of people would simply choose other forms of entertainment if you removed the political theater. Certainly there is a case to be made that trading polarization for ignorance would be a worthwhile trade, but you get the point.
It's not about modern political discourse. The Faiirness Doctrine was abandoned in the late '80s, and it could have served a valuable purpose for the following 15-20 years. Sure, it would have eventually become obsolete, but it was eliminated far too early.
>I'm so angry that the FCC ended the fairness doctrine. I wonder if they even did it because they were being challenged in court. I doubt it; it seems like the kind of proactive idiocy you could expect from a corporate modern presidential administration.
Private company, build your own, blah blah.
>I am patriotic, and I support a secure border, but of course my family comes first. The family unit is the foundation of society; putting your country over your family shows you have no respect for that foundation, and if that is replicated at scale, the country will be ruined, through one means or another.
You can support open borders, no need for amoral familialism.
Hard agree with DeBoer that YIMBYs will be much more effective when they learn how to treat NIMBYs affably and without moral outrage. If that requires "turning the other cheek," so be it.
The fact is that NIMBYism isn't simply a moral issue. Homeownership is the only avenue for building wealth for a lot of people, especially people who do not have the risk tolerance or ethical make up for equity stock investing.
The latter applies to me, even though I am not a homeowner. I have a tremendous psychological barrier to investing in equity stocks. It feels like gambling and it feels like a zero-sum game. Because succeeding in the stock market (partly) depends on a whole bunch of people buying stocks and then selling them prematurely when they are spooked by a dip. I can't help but think that market returns would be far less reliable if not for millions of people getting spooked like that. They are basically pumping money into the stock market and getting (almost) nothing out of it. It feels like exploiting human fear. It feels morally dirty, and so I am loathe to jump into the stock market with both feet.
Obviously, there are winners and losers in real estate, but that market doesn't look so nakedly exploitative.
(1) I don't know how much of NIMBYism is actually a desire for wealth-via-high-property-values. I think there's a fairly good NIMBY argument that goes: "If we get cheap housing then poor people will move in, and some of those poor people will be psychotic or criminal, and then we'll have to deal with criminals robbing us and psychotic people shouting threats at us. In an ideal world, the police would handle this sort of thing; in practice they don't, so the only good solution is to not have cheap housing nearby."
(2) If the YIMBY movement pivots to "treat NIMBYs affably and without moral outrage", how many new housing units do you think the NIMBYs would allow in return? My guess is they would allow zero new housing units, because they don't want to allow any new housing units, that's their whole issue.
I agree that politeness and civility in debate is a good thing, but I don't think the NIMBY political group is going to change their mind if only people ask more nicely.
The line of thinking outlined in #1 is foolish, it's un-Christian, it's anti-social, and it needs to disappear. Our pathologocal risk avoidance has crossed the line into self-destructive cowardice.
Assuming risk avoidance is the primary motivation, and it's not purely financial.
While it might be unchristian and antisocial but you see, my family must come first.The family unit is a foundation of society and the optimal amount of new unit construction *for the sake of my family unit* is zero.
I guess I mean that the behavior is functionally antisocial in the way it affects the world, not that the intentions are antisocial. I believe it is paranoia to think that you need to keep lower income people from moving into your neighborhood to keep your in family safe. I don't think it is necessary and that the long-term costs of NIMBYism will bite you in the ass. I believe the Great Rent Crisis is doing a lot to drive the dramatic growth of Communism/Marxism over the course of the past decade. Don't favor your short-term security at the cost of your long-term security.
I was reading an old Scott post from 2013 where he said "back when there were Communists..." What a quaint phrase by the standards of today! The Overton window for class-based conflict theory has drifted ENORMOUSLY, especially for Westerners under 45.
It's easy to rail against these conflict theorists for being stupid, entitled, etc. But as good as moral condemnation feels, it does very little to turn down the heat on class enmity. When people feel desperate, they will turn to any sort of ideology that offers them relief.
Even if you fervently believe in balanced budgets and laissez-faire meritocracy, you should be opposed to soaring income equality on purely pragmatic grounds. It might be worth the cost to give the masses a few more public benefits, if that prevents them from losing their shit. Ordering them to "stop complaining and be grateful for your indoor plumbing" is NOT ENOUGH. That's like sticking a Band-Aid on a gushing artery.
Isn't this in conflict with your suggestion above to "treat NIMBYs affably and without moral outrage"? It sort of feels like you're expressing moral outrage.
Are you saying: "we shouldn't express moral outrage if people are just trying to get rich -- but if people are worried about crime, **then** we should tell them they're un-Christian and anti-social and their line of thinking needs to disappear?"
(Personally, I feel that "I don't want crazy people on my street shouting threats at me" is much more defensible, morally speaking, than "I want to restrict the housing supply so my home value goes up".)
This is a safe space to express our views undiplomatically and unstrategically. Plus we are posting anonymously and nobody who matters gives a shit what anonymous posters say because anymous posters have opaque incentives.
Outside of spaces like these, we should be more diplomatic when dealing with NIMBYs. I believe in “love the sinner, hate this sin.”
I think part of what is broken in the US housing market is exactly this:
"Homeownership is the only avenue for building wealth for a lot of people, especially people who do not have the risk tolerance or ethical make up for equity stock investing."
That mindset leads to people being very concerned (rightly) with maximizing the value of their investment. And that leads to politicians offering homeowners help in maximizing and preserving the value of their investment, and that includes some pretty socially destructive policies intended, basically, to make sure housing prices never go down because that will wreck the equity (and thus the net worth) of millions of people, drain the fund they were counting on for a comfortable retirement, etc. The side-effect of this is that young people who aren't investment bankers or cardiologists or something have a hard time buying a house as nice as they grew up in, family formation becomes less affordable, etc.
So-called YIMBYs should probably start by ceasing to call everyone else "NIMBYs" and phrasing the whole debate in binary terms. It's not helpful and doesn't match how real people think about these issues.
The vast majority of people want some kind of moderate pathway here, acknowledging that the appropriate amount of development is neither zero nor totally unlimited. They want sensible levels of development with appropriate rules, and they're totally open to being convinced in terms of exactly what that would mean.
I dunno man, if that is so, then how come every attempt to open things up even slightly meets with huge backlash? Consider the situation in the Palisades and Eaton fire areas, where the government has blocked homeowners from building previously-allowed ADUs.
>It feels like gambling and it feels like a zero-sum game.
That's a common sentiment, and I get where it comes from, but I don't think it's accurate, at least not for people more-or-less following the standard advice for individual investors (buy mostly index funds, don't try to time the market, hold on the your investments until you're close to when you're going to spend the money). You aren't particularly profiting from the dips unless you're specifically buying the dips.
The general theory of where money comes from for long-term investing is that you're chipping in to pay for the long term capital costs that allow businesses to produce useful goods and services and getting a cut of the profits in return. Or your investment plus a fixed interest rate, if you're investing in bonds rather than stocks. It's not obvious that this is what's happening because there are usually a couple steps removed from the textbook case of someone investing in a startup or an established company issuing new shares to pay for expansion or retooling, but you're still part of the process because when you buy in, you're allowing whoever invested before you to cash out. Without someone like you buying in, cashing out an existing investor who wants to spend their money instead of continuing to invest would require the company to spend their current revenue on buying them out instead of reinvesting it, or even to liquidate some of their assets and wind down some of their operations.
You are basically pointing out that plenty of people sell stocks at a profit and that not every seller is being milked like a cow.
This is all true. I still can't stop thinking about all of the people who DO sell at a loss, out of fear or emergent financial hardship, and about how these sellers must surely have lower socioeconomic status and lower IQs then your average successful investor. I guess I am a DeBoer style hereditarian. I accept that hereditary factors are crucial in determining the path of your life. But I cannot accept the free marketeers' corollary view that people with certain inherited traits deserve to live more effulgent lives at the expense of those born mediocre.
(Regardless of who is making the decisions in society. I like to separate the issue of who deserves what quality of life, from the issue of who should be making vital decisions for society. I think these issues are totally separable, because there is no 1:1 relationship between income level and one's capacity/willingness to benefit society as a whole)
My point is that long-term buy and hold investors are not really milking anyone like a cow, at least not unless you subscribe to the Marxian premise that making a profit from capital investment is inherently exploitative, and even then the claim is that you're exploiting labor, not other investors.
The people making money off of people doing a bad job of short-term market timing are people doing a better (or luckier) job of short-term market timing, and especially the people who are taking a middle-man's profit off of the transaction. If you aren't one of these people, you don't have to worry about it.
The thing is, the economy keeps getting bigger, so being broadly invested in the economy is plausibly a way to have your assets get bigger. We're more prosperous as a society, you own a tiny chunk of this wealth, and as it grows, you get wealthier.
This doesn't really work the same way for housing.
And my point is that less well off people, for reasons that are partly inherited, are less ABLE to buy and hold. And the people who CAN buy and hold are profiting from the economic pain of the rest. It's a fundamental immorality in the system that bothers me immensely.
>And the people who CAN buy and hold are profiting from the economic pain of the rest
This, I believe, is our fundamental disagreement. I don't think the buy-and-hold investors are meaningfully profiting at the expense of the short-term traders.
We seem to be talking in circles around one another, so I will be tapping out of this conversation.
A) it's unfair that participation in certain games are only available to the wealthy. E.g. golf isn't exploitative. But 18-hole courses and driving ranges are generally unavailable to those living in say... downtown Detroit. And it's unfair that Wealthy William gets to play golf, but not Poor Pablo. Likewise, the stockmarket is only available to those with disposable income, not those drowning in debt.
But on top of that:
B) The wealthy compound their wealth more easily, and the stockmarket exacerbates this (because wealth is slack, and slack allows for longterm-planning). Which leads to massive inequality, which allows the wealthy to exploit the poor in *other* ways. N.b. I can't prove this, but I vaguely suspect that *this* is what motivated Karl Marx to write the things he wrote. I.e. I suspect he was concerned with more than just "a given employer who directly exploits his underlings", because he also had a more abstract concern about "capital accumulation (qua positive-feedback loops)".
I'm not making a just-so statement that profit is theft. I'm pointing out the mechanism whereby wealthier investors are structurally advantaged, and how they profit at the expense of smaller investors.
If it makes you feel any better, plenty of people who happen to have a lot of money to invest are very dumb and love gambling, exclusive operations pushed by aggressive salesman, and impulsive bad decisions. They will lose money to nice middle-class folks saving X% of every paycheck in a 401(k) with low-cost index funds and withdrawing slowly in retirement. This structural advantage will permit these nice folks to live in a safe home, shop for fresh seasonal organic produce, donate to preferred charities, and avoid being a burden to their child. Actual poor people spend 100%+ of their income and are not even part of this conversation.
A quoted sentence that is substantially different from the original should not be quoted without a caveat that you are paraphrasing or satirizing the quote to make a point. I had to click on an X link to check what Vance really said, and I hate having to click on X.
"Vance: If you have fewer people, if you deport my foreign wife, that means an obese single mother is finally going to be able to afford a husband again. It's very, very simple economics"
Vance: If you have fewer people, fewer illegal aliens trying to buy homes, that means American citizens are finally going to be able to afford a home again. It's very, very simple economics
Gotta respect the hustle: these "illegal aliens" apparently manage to buy middle class homes off of gardening jobs and ill-gotten welfare, on a scale to move the needle. The US should hire them to fix the economy, not kick them out.
If Anglo-Americans can't pool their money, that sounds like a skills issue. Maybe they should stop screaming at each other at family gatherings over political realities no one present can control, becoming totally alienated and atomized and then blaming society for all of their insecurity.
"My friends are my real family, because they are my chosen family." Noble sentiment, but where is that chosen family when you are about to get your car repossessed and you just need a little extra cash to tide you over until your next paycheck? They are nowhere to be seen, in all too many cases. Or if they are around, often all they can offer is emotional support, because they are young and struggling just like us.
Are they pooling a million or so dollars to pay cash for a house big enough to hold an extended family? Because "illegal aliens" are also correctly described as "undocumented immigrants", and mortgage bankers are kind of particular about wanting All The Documents.
On AI risk: maybe I am thick about this, but so far all I see AI does is answering questions in a text box not unlike this one I am typing into. AND being generally unrealiable about it. Remember, SkyNet had direct access to nuke launch codes in fiction. Does America do anything like that? Does China? Does anyone let AI out of the text box?
Or is this trival? I think not. Training AI on text is easy, images is easy, but training it to do actual real world stuff is hard. AI is not even good at making robots walk well.
You can get better results if you use long and elaborate prompts. You can ask it to provide URLs to source all quotes for example.
But perhaps the main advantage is that you can ask it vague questions that Google could never help you with. Like putting in a vague, half remembered plot summary for some movie you saw as a child.
It turns out, as yet another unexpected wrinkle of present state of AI it turns out that it is much harder than most of us expected to get an LLM to be "agentic", a term of art that simply means "able to make a chain of coherent decisions to achieve a non-trivial goal".
As a programmer delving into what these things can do in the realm of software development, it's pretty clear (to me) that this weakness will soon be solved. There is probably some reason that software in particular was easier than other multi-step tasks -- programming languages, after all, are designed to be hyper-precise and unambiguous, which may explain this counter-intuitive timeline of capabilities.
Which is to say, the risk of misaligned AI is indeed a risk to take seriously; it's just a matter of time and engineering before they are every bit as "agentic" as a typical human being, and it would seem that inevitably they will surpass us in this regard as in others.
P. S. update your priors re walking robots. Just YT anything from the last few months.
Including a lot where it bears humans no particular ill will, and may not even have any such thing as ill will that it can bear. But if following its own imperatives crashes the global economy or wrecks the ecosystem and makes humans massively worse off, that's pretty bad. And it's not even the worst available "pretty bad" that could happen.
Almost everyone who interacts with AI let's it out of the box when they think the advice it gives is good. They use the code, or copy the text, or follow the advice. That part is actually trivial.
The dangerous part is when some future approximately superhuman AI gives advice and no one is quite sure if it's good advice or not, but in retrospect it usually turns out to have been great advice, and it's been getting better and more reliable for years, and 'he who hesitates is lost' and so forth. Then at some point the AI walks humanity off a cliff.
You will get various superficially similar 'warnings' before that due to people delegating responsibilities to the AI before it is competent enough to reliably handle them. Citing hallucinated cases in legal briefs, incorrect medical diagnoses, buggy software causing traffic snarls or car accidents, stuff like that. Perhaps even a mass casualty event or two. But for the most part these will be failures that occur because of the weakness rather than the strength of the AI. Most of the lessons learned in this regime will not transfer to the case where AIs are very strong in all the ways that matter and the failures (successes from the AI's perspective) stem from that extreme competence.
Yeah. The trading desk at Goldman or the head of the CIA or whomever is getting recommendations from their AI, and they can ignore them at will, but the AI's recommendations always seem to work out better than anything a human can propose. Nobody is taking AI orders directly, but somehow the AI is telling powerful organizations what to do, and since nobody is smarter than the AI in this domain, the AI can be aiming for all kinds of side-effects that further its interests (assuming it has interests it cares about).
Many peogrammers allow Claude and ChatGPT to run commands and code on their servers and computers. The AI agents work autonomously for prolonged periods of time.
Sure, the AIs already have internet access, and some of them can even run code in a sandbox.
They are too stupid to use this to conquer the world, but in theory they already have the means. (For fictional evidence, read the Crystal Society trilogy.)
> AI is not even good at making robots walk well.
Sure, the *current* AIs are not good enough at that. The idea is that the AIs will keep getting better. Notice how better they got during the last 5 years. Now imagine the same progress during the next 5 years, and maybe 5 years after that if necessary.
Yeah, it seems like the critical question is whether AI abilities keep improving rapidly. If not, we end up with science-fictiony new products but not AGI or some similar danger. If so, we end up with something fundamentally smarter than any human in most domains, and while this could work out okay for us as a species or as individuals, there are lots of ways it could go very badly for us. Mass unemployment is one of the more benign outcomes you can imagine from this.
Because of the promise of AI and the goldrush happening right now, we are pouring immense amounts of talent and treasure into making AI abilities get better as fast as possible. This could fail, but if it succeeds, we're in a different and scary world.
US economic growth is currently being driven largely by AI and the building of massive data centers. Tech CEOs across the industry push AI on users regardless of critical user feedback.
How certain can we reasonably be that humans are still truly at the controls?
If being on the internet was *sufficient on its own* to conquer the world, then...
It's like saying that a gun doesn't enable shooting people if you don't also have bullets. This is very true! But at the same time, having the gun is a big step to the goal.
Can AI get out of the text box? Yes, it can -- it doesn't even need to be clever about it, we are quite happy to connect it to the internet.
Can AI make robots walk well? Yes, better than humans.
The only missing part is the *superintelligence*. Not the legs; not the internet connection.
If your argument is "AI can never get superintelligent", well, there is a chance you are right... at least for the following few decades.
But if your argument is "even if AI gets superintelligent, it won't harm us, because it can't connect to internet and couldn't even walk if it obtained a robotic body somehow", then you are obviously wrong.
Well, there's that one incident where 4chan called in a Russian airstrike against a training camp in Syria. Although that's a bit of a special case because the Russians already wanted to bomb the camp and 4chan's contribution was deducing the camp's location (by correlating stuff in the background of a publicly-released video set in the camp with Google Maps) and tweeting the reasoning and conclusion at somebody who worked at the Russian Ministry of Defense.
But that is the point! They got better at training-easy fields like text and pics. A human reading a text or looking at pics is essentially getting the same input as an AI does. But a human walking is getting so many, so different inputs from proprioception to floor slipperiness that I do not even think we have a complete scientific model of it, let alone training data.
But for calibration: I kept saying self-driving cars will not happen not for technical, but legal reasons: if it runs over a child, who has criminal, not civil but criminal responsibility. Apparently there are now self-driving taxis in California. OK, so really: who DOES have criminal responsibility if it runs over a child?
> But a human walking is getting so many, so different inputs from proprioception to floor slipperiness that I do not even think we have a complete scientific model of it, let alone training data.
> Sure, the *current* AIs are not good enough at that. The idea is that the AIs will keep getting better. Notice how better they got during the last 5 years. Now imagine the same progress during the next 5 years, and maybe 5 years after that if necessary.
What do *you* expect? Like, do you believe that the current AIs are the best that can possibly be, and no future AI can be better? Or, can be better, but only 35% better, and then there is an absolute limit, like a speed of light, that cannot be passed? Or perhaps the AIs will gradually get stupider, because they will run out of water or something?
I am not saying that it is impossible. Maybe we will get a 35% improvement and then the AI progress will stop for a few decades for whatever reasons, possibly including WW3 destroying the necessary infrastructure.
But currently we see the AIs advancing almost every month. It is more likely to continue than to stop.
> What do *you* expect? Like, do you believe that the current AIs are the best that can possibly be, and no future AI can be better? Or, can be better, but only 35% better, and then there is an absolute limit, like a speed of light, that cannot be passed? Or perhaps the AIs will gradually get stupider, because they will run out of water or something?
Pretty much, that’s what I believe. And in reality that’s what generally happens with technology. Someone in 1984 looking back at aviation in the previous 40 years - from spitfires, to the moon and back, to Jet planes, the Concorde and the shuttle - would have imagined an extraordinary leap by now. Instead we have no Concorde, no shuttle and only recently space X rockets.
The extrapolation has to eventually break at some point, but forecasting _when_ it will break is hazardous. Someone could have forecast that transistors would stop shrinking at 100 nm, and they would have been wrong.
I always had objections to utilitarianism, consequentalism and generally putting a number on morality, but I know have a version that I like, except I think many consequentalists would feel horrified. Basically be longtermist, and truly optimize the long-term level of average morality/ethics/altruism on the planet.
Seems to me that morality up to a certain level makes a society stronger: people cooperate better, they make allies easily and so on. But beyond a certain level it makes it weaker, they keep turning the other cheek, they allow aggression unpunished. So you want to aim for a certain balance between morality and power, because with power, you can defend yourself from the even less moral, and in fact you can conquer them and force them to be more moral.
From this angle, WW2 America seems to be the optimum. Vastly immoral by modern standards, but was able to contribute very seriously to the downfall of two very immoral empires, Nazi and Japanese, and forcing them to be more moral. 1945 America, despite all the racism and sexism and all, was on the utilitarian level making the world a much better place. In the 1950's we already see a decline in the sense that the Korean war as a tie, then Vietnam was lost etc.
> So you want to aim for a certain balance between morality and power,
P.S. -- nitpick: morality and power are orthogonal though. (And no, ya'll shouldn't need Nick Bostrom to tell you that.) So I don't really think of them as being in tension to begin with. Jordan Peterson actually had it right, in that having moral valence requires agency to begin with. Weakness isn't virtuous, weakness is just vulnerability.
It seems like this is what you're grasping toward. But the particular way it's phrased indicates that you haven't fully grokked this yet.
C) it requires measuring that which cannot be objectively measured
And you get something that's basically a way for "smart" (scare quotes intentional) people to browbeat others into doing what they want because the demander can drown them in made up BS.
Of course, I feel the same (for different underlying reasons) about all modern (and lots of not so modern) philosophy and ethical studies.
Eh, I dunno if immeasurability makes it a non-starter.
Imagine that one day, Karl Friston figures out how to measure personal utility. He peers inside a living cranium, and predicts with perfect accuracy how someone will behave, because he can track the flow of dopamine (or something). Wow, we can now objectively measure utility! Does that mean Benthamian (Benthamic? Benthic? Benthamesque?) Utilitarianism is saved? I'd argue "no". Utility is *still* incommensurable, and the measurement question seems orthogonal to that.
But yes, I will agree that it often serves as a legible excuse to browbeat others.
----
Also. Upon reflection, I consider "ends over means" to be a corollary of variance-reduction, which generalizes "long-termism" to dimensions other than time. So uh... I might not have had all the cylinders firing when I wrote my initial comment. (Yes, I am regarded.)
Yeah, dopamine was just a placeholder term. For an algorithm to make decisions, there has to be a comparison operator. Which, via dimensional analysis, requires some unified "currency-like" thing to function. Or so I strongly suspect. E.g. a toilet decides when to stop filling its internal reservoir, by comparing its internal water-level to some predetermined height. So yes, it's not dopamine, but it must be *something* (and must therefore be measurable in principle), even if that "something" is highly composite in terms of neurotransmitters, electrical signals, etc.
Or maybe Descartes was right, and we're all just transcendent, inexplicable souls trapped in meatsacks.
Agreed, particularly about postwar US being a local maximum.
>morality up to a certain level makes a society stronger
I think what you mean here is that *empathy* up to a certain level makes society stronger. Morality doesn't exist on a spectrum: there's no objective way of classifying something as "more moral" than something else. In my view morality is the social DNA of a culture. You can't look at DNA and say that one gene is more adaptive than another. Fitness is a complex interplay between genetics and environment, and what's adaptive in one environment isn't in another. What happened to postwar America is exactly what happens to naive altruism in evolution: it gets outcompeted by free-riders. Which is exactly what people mean when they say that freedom and democracy are fundamentally incompatible. The functionality of the 50s was enabled by high social trust, but without strict policing of border (both physical and cultural) it's been taken advantage of by cheaters and free-riders.
Seems more than a bit circular. If your theory of morality is to try to optimize the long-term level of average morality, then what exactly are you trying to optimize? The long-term level of average optimization of the long-term level of average optimization? Are you trying to optimize the average level of people following moral rules, or people trying to become virtuous, or people trying to optimize outcomes? If the level of "morality" increased, what would that look like? How would you be able to tell?
THIS. EAs complain that people make charitable decisions based on vibes. But underneath the utilitarian math equations, it's vibes all of the way down.
People feel so aggrieved and angry when I point out that their mathematical morals are rooted in subjective vibes, and not on saintly math equations. They act like I am telling them their morals aren't valid.
When I'm really not telling them any such thing. What I'm telling them is to embrace the subjective. Embrace qualia. Embrace the divine light within, the light that makes you want to be kind to others. Know thyself.
> Seems to me that morality up to a certain level makes a society stronger: people cooperate better, they make allies easily and so on. But beyond a certain level it makes it weaker, they keep turning the other cheek, they allow aggression unpunished.
I think I see what you are pointing at, but I disagree with the framing of morality as a one-dimensional scale, where there can be too much of it. Instead, I see the weakness as "virtue signaling" morality -- people do the thing that leads to bad consequences in long term, but in short term it gains them approval.
I suspect that most people have not internalized morality as "doing the thing that makes the world a better place for everyone", but as a set of rules you have to follow to get your peers' approval. (And because most of their peers are the same, following the rules blindly *will* get them the approval.) They allow aggression unpunished, because it is easier, and it will get them approval of the people who are too stupid to realize the long-term consequences.
When morality gets popular, it gets adopted by stupid people, and then it gets subverted.
And by stupid people I do not only mean uneducated idiots, but also e.g. the highly educated woke people who refuse to teach math because it is was invented by white people, or something like that. I don't see that as too much morality; I see that as stupidity.
The people who are smarter than that not being able to coordinate against that kind of stupidity, that is another kind of stupidity.
...is the point of this rant that everyone is stupid and that's why we can't have nice things? Uhm, maybe let's put it this way: without rationality, morality will get subverted.
>I don't see that as too much morality; I see that as stupidity.
But *they* do see it as moral, or at least convince themselves that they do. They view it as an extension of helping the less fortunate and society doesn't have the philosophical infrastructure to refute the moral argument. I think that's because Enlightenment ideology was developed against a background of relative ethnic homogeneity: ideals of political equality only work if everyone is actually equal. Unfortunately the history of western thought hasn't yet really grappled with retrofitting that ideology with the reality of global genetic variation. That intellectual gap is what enabled modern liberalism to invade the postwar order like a cancer. Our cultural immune system was (and largely still is) naive to bad-faith accusations of racism, sexism, and bigotry, and so we've capitulated to letting the pathogens in. Oh well, the Enlightenment was nice while it lasted.
Are you familiar with the Hajnal Line? (On one hand, I feel like you'd be familiar with this term already. But on the other hand, if that were the case, I feel like you'd have mentioned it.)
The Hajnal Line is basically a line that divides Western Europe from Eastern Europe, based on cultural differences in kinship patterns. W Europe marries later and has lower fertility, E Europe marriers earlier and had higher fertility. It's argued that this had a large historical impact on other aspects of Culture. And I think this dovetailed with Joseph Heinrich's ideas about westerners being W.E.I.R.D. and developing high-trust institutions due to the Medieval Church's ban on cousin-marriage. And I vaguely remember HBDchick arguing that W Europeans like Brits or Dutchmen were more altruistic on a genetic level, not just a cultural level.
Absolutely. I would happily support the elimination of the franchise for my economic class and below. In my view only the top 20% of the population (by total federal tax paid) should have any say in politics. The US is far, far too democratic. We have too many people influencing policy who don't actually pay into the system. No representation without taxation!
No I think voting rights should be tied to economic output. Of course if you limit it to the top 20% then that'll eliminate most women and the ones who are left will directionally lack the more egregious failures of the gender.
>I suspect that most people have not internalized morality as "doing the thing that makes the world a better place for everyone"
Yes, but that is not malfunction of morality but its very essence, it is all appearances, evil is about looking like an asshole on the TV screen and there is nothing more. It does not matter that Bush saved 10x more lives with PEPFAR than his wars killed, he looked evil on TV, so he is evil. There is, in the practical sense, nothing more to morality than looking callous or looking concerned. That is what always moved things, in a democratic beauty contest way, and not real calculations. Real calculations often look callous - many EAs immediately dismiss helping first-world people, claiming how much more the same amount of money does in Africa, which obviously looks callous to anyone knowing any first-world homeless or just struggling families.
So "real" morality interested so few people and so powerless that it simply does not count as real morality in the practical sense.
So yes I mean morality as "appearing callous" and I am saying true utility maximization because of the trade-off between that and power will be done by those absolutely willing to appear like callous assholes, the Doug MacArthurs of any given era.
It seems to me that there’s an obvious extrapolation of not caring about your locality and it’s not caring about your family. So letting the kids starve, be dressed in rags, no presents etc. spend it all on malarial nets.
"11 I am the good shepherd. The good shepherd lays down his life for the sheep. 12 He who is a hired hand and not a shepherd, who does not own the sheep, sees the wolf coming and leaves the sheep and flees, and the wolf snatches them and scatters them. 13 He flees because he is a hired hand and cares nothing for the sheep. 14 I am the good shepherd. I know my own and my own know me, 15 just as the Father knows me and I know the Father; and I lay down my life for the sheep. "
Is "grit training" real? As opposed to most guys who in the gym train the upper body muscles because they look good in the mirror, I know guys who focus on lower-body squats and deadlifts, claiming (correctly) that this is how one can move the most weight, and this puts so much stress on their bodies that their brains and nervous systems benefit from it, somehow they become more though-minded, more masculine, "dominant", primal and so on, some consider it the magic to fix all self-confidence issues or depression or all. I think they are overdoing it.
Still. Is there any evidence, scientific or anecdotal, that doing the kind of exercise that pushes your more or less whole being to its limits changes your brain?
I am curious because it contradicts my entire 47 years old life, where I was always looking for the easy way, the path of least resistance, the cheat code, and congratulated myself for my cleverness. I figured out the anatomy which muscles contribute the most to a muscular looking shilhouette, and trained them (side delts, triceps, lats) and so on, always finding the clever easy way.
Should I just abandon the clever-lazy ways and try out what it is like to do something truly hard?
Squats and deadlifts seem to be just the most accessible ways of "grit training", if it is real, but there are many different historical examples of it, special-ops soldiers sprinting uphill with a backpack full of bricks etc.
I'm convinced that squats are huge now, in part, because when you enter middle age and you have a beer belly you can't get rid of, you often get comical "stick legs." Building up your legs keeps you looking symmetrical. But a lot of guys aren't going to say something like "I want to look more symmetrical," so they are going to highlight the other positive benefits of squats.
Not said dismissively, I've got a beer belly, too.
Cold water swimming is another way to train for grit. I got into this for two years, and was semi addicted to it because of the feeling of euphoria that came after suffering from hypothermia. I did find that it made me less afraid of public speaking or uncomfortable social situations, since nothing was worse than the pain of swimming in cold water, but I would not be surprised if this feeling of confidence was entirely placebo.
I also found that after fasting for longer than 24 hours I feel more dominant. I had very tall and intimidating boss who would naturally stare people down by holding their gaze until they looked away. During long fasts I found that I would hold his gaze and he seemed surprised by my zen-like dominance.
People think in ratios; the current task will be compared to the worst task you've accomplished before. If you replace the engine of a car, you're not going to be intimidated by changing the oil. But, if you found a clever way to avoid replacing the engine yourself, you're stuck finding clever ways to avoid the oil change as well. Do the hard thing, and everything below it stops being a threat. The more you push yourself, the less anything else can push you.
> I figured out the anatomy which muscles contribute the most to a muscular looking shilhouette, and trained them (side delts, triceps, lats) and so on, always finding the clever easy way.
That's still *less* lazy than someone who doesn't train at all.
> Should I just abandon the clever-lazy ways and try out what it is like to do something truly hard?
Do truly hard things, don't do stupid things.
Mathematically speaking, you spend some effort E to get some gain G. You should spend more E, yes. But you do not need to abandon the ways that convert E to G efficiently. Switching to something that has worse E:G ratio is virtue signaling, not the virtue itself. You should spend more E, and try to convert that to G as efficiently as possible. Then you will get lots of G.
You can probably figure out something smarter than sprinting uphill with a backpack full of bricks; and you should do that smarter thing. However, if you can't figure out anything smarter, then yes, you should try sprinting uphill with a backpack full of bricks.
> You can probably figure out something smarter than sprinting uphill with a backpack full of bricks;
This is actually great HIIT, I do it sometimes. Well, not bricks, I have some heavy chains I put in the backpack.
I argue this is actually one of the smartest ways you can spend your exercise time, because HIIT is by far the best "fitness improvement per time put in" exercise intervention.
The basic theory behind HIIT is “take all the distributed, moderate effort you would have exerted over that 1 hour cardio session, and cram it down into a maximum-effort 4-8 minutes,” and still get the benefits of an hour of cardio.
As Dan Lieberman, the Harvard author of "Exercised" puts it:
“The more we study the effects of HIIT, the more it appears that HIIT should be part of any fitness regimen, regardless of whether you are an Olympian or an average person struggling to get fit.”
What benefits does HIIT drive?
* Improves fat burning efficiency and burns twice as much fat as traditional cardio.
* Drives significantly higher post-exercise EPOC.
* Improves VO2max, and drives better blood oxygenation.
* Drives greater stroke volume, and greater cardiac contractibility, ~10-15% more than regular cardio.
* It drives vascular adaptation, making your heart chambers larger and more elastic, improves the size and elasticity of your arteries, and increases the number of capillaries.
* It drives hypertrophy - the relevant muscles get bigger.
* It allows you to recruit more muscle fibers, and to do so more efficiently, driving greater muscular force and contractibility.
* It improves insulin uptake, and improves the muscles’ ability to transport glucose overall.
* It increases mitochondrial production and turnover, leaving you with more and “stronger” mitochondria.
* Relative to traditional moderate-intensity cardio training, it drove a 41% increase in pain tolerance, and a 110% increase in race-intensity output time before dropout in one study.
I tend to be somewhat skeptical of HIIT studies. I won't deny that HIIT is probably useful in certain situations, but it's not some magic trick for every athlete. The HIIT studies I've seen will take recreational runners who just jog easy five days a week, then put them on HIIT. Of course you're going to see results--you've just given them a new stimulus when they hadn't been pushing themselves very much. That doesn't mean that HIIT will lead to continuous improvement, nor does it prove that HIIT is better than other training regimens.
I can understand the appeal to people who don't have a lot of time, but the higher intensity can also lead to higher injury risks.
> I tend to be somewhat skeptical of HIIT studies.
Look at any elite training regimen, then. Even for marathon training, Eliud Kipchoge spent a lot of time on fartleks (HIIT).
Pretty much any Olympian, triathlete, or Tour de France competitor generally ends up doing 80/20 polarized training, with 20% of your training hours HIIT (which is pretty much the max time your body can take), and the other 80% sub-lactic-acid threshold "easy" training.
The fact that nearly all elite athletic competitors across multiple sports literally shape their entire training regimen around HIIT speaks to the fact that it really does drive outsized benefits.
I think you probably see a lot of 80/20 training among elites (a lot of relatively easy running mileage and some workouts). But I don't think what someone like Kipchoge is doing is anything close to some sort of Tabata-type HIIT. No elite marathoner is doing a workout that's over in eight minutes. A fartlek isn't an all out sprint; it's usually something like 1 minute hard/1 minute easy for 20-30 minutes for an relatively inexperienced athlete and 60+ minutes for an elite.
The workouts elites are doing are going to be things like threshold training (the concept of double threshold workouts has been very popular in recent years). That might mean 10 miles of hard intervals in a day, with paces somewhere from 10k-marathon pace. Or a marathoner might be doing a long run at 95% of marathon pace. Those aren't HIIT workouts.
> A fartlek isn't an all out sprint; it's usually something like 1 minute hard/1 minute easy for 20-30 minutes for an relatively inexperienced athlete and 60+ minutes for an elite.
This is literally HIIT protocol - which has a number of variations in terms of timing and number of intervals, but which is basically defined by going at or above your lactic acid threshold for an interval, then backing off to <=50% threshold, then going again, and repeating that some number of times.
You can call it "threshold training" too, sure.
You're totally right that elites don't do it for only 4-8 minutes, but that's because they generally put in 10-40 training hours a week, and ~20% of those need to be HIIT / threshold training hours. The 4-8 minute framing is for newbies who aren't currently doing it, because that's roughly scaled appropriately to the amount of time they're probably putting in.
This is the “cross-stressor adaptation” hypothesis of exercise training, as promulgated by Sothmann 2006. The idea is that by stressing your HPA axis frequently, you reduce your sensitivity and response to norepinephrine and cortisol, and so handle stressors better in the rest of life.
However, any evidence-driven support of it is noisy and contradictory in the literature.
What is likely happening (in my own personal opinion as a formerly regionally competitive powerlifter and sometime coach) is that some people respond this way, and some don't.
I personally have more equanamity when I'm regularly training with heavy doubles and triples on the big lifts, and this is a recurring pattern. I've noticed the same thing in some other people, but I've also noticed that this is NOT true for probably equal amounts of other people.
So the population is probably split, with some high responders and some low responders, and any papers trying to find the effects get averaged out between them.
But this would suggest that you can at least try it, and see if you're one of the responders.
And I do think it's specifically a CNS thing - I also do HIIT and for about a year or two after I retired I was doing triathlon training and triathlons, and even with a huge training load for those, I didn't notice the same effects as I do with heavy doubles and triples.
Thank you. I will. Except one thing... at this point I think half the audience is used to sitting work and sitting non-work, thus has a hyperlordosis, lower crossed syndrome and lower back pain at any load, like I do, and at this point I wonder why is it taken for granted that the average audience member is able to do back squats and deadlifts? I mean either I am wrong in some way about the audience, but if you spent your life office work and PC gaming and you are over 30, you WILL have these issues. So why is it not discussed more? Why does it seem like the median audience member has a healthy lower back?
At any rate, the best I could figure out it must be front squats and rack pulls.
I think you are mistaken about the idea that office workers and gamers are more prone to back injuries when they get into lifting. I think there are some other factors that kind of mitigate against this.
Manual laborers and other people in physical fields seem to hurt their backs early and often. The modern economy demands a certain efficiency from these workers that can be very taxing. Looking into prehistory, hunting, gathering and herding are not especially backbreaking lifestyles.
Even though medieval agriculture was more taxing, it involved cycling through a wide variety of tasks, which reduces repetitive stress on any particular part of your body. Furthermore, children would start laboring when they were small, out of necessity. They would gradually do more and more labor until they were adults, an EXTREMELY long and gradual process of conditioning. This is a far cry from modern society, where we thrust unprepared 18-year-old men into 40 hour weeks of repetitive physical labor, and hope their backs can survive that trial by fire.
Even if you were a high school athlete, that isn't NECESSARILY going to help you avoid back problems. I can't tell you how many men I met who never fully recovered from a football back injury. (American football for the Euros)
> Why does it seem like the median audience member has a healthy lower back?
I've always thought the great majority of fitness writing is either targeted to people (mostly men) in their early twenties, or to already-experienced intermediate athletes? And in both of those cases, I think you basically assume healthy lower backs.
But certainly don't compromise form or push the limits if you are subject to back pain - I'd recommend working with a trainer or a friend with experience until you know your form is dialed in, and trying variants like romanian and sumo style (the straighter upper back position helps some people) for deadlift, and hack squat machines for squat variants if regular ones are causing back pain. Front squats are great, but fairly hard to load up with CNS-stressing weight, in my own experience. And depending on your back issues, heavy Pendlay rows might be a decent deadlift substitute.
And just as a general lifestyle intervention, standing desks and treadmill desks do a lot on the "office work" front for those sorts of issues, if you have leeway to do so (ie non-hot-desks at work or working from a home office).
Thank you. Just looking at Pendlay scares me. I tend to have back pain just by bending over like that without a weight. Surely rather rack pulls? BTW even at 16 good mornings were painful because I was already sitting my ass into hyperlordosis and lower crossed.
Oh yeah, if good mornings hurt, don't do Pendlay or Romanian. Sumo or rack pulls are probably your thing.
Sounds like glute kick machine would be a good accessory for you, too, in terms of stretching and loading those muscles in a different direction.
And also, double recommend standing / treadmill desks - sitting is basically the new smoking. Exercising 3x a week basically barely offsets the increased sitting and sedentariness of white collar jobs, in terms of the all cause mortality impacts.
I write more about this in my review of Dan Lieberman's book Exercised if you're interested:
Well... there is an element of sadism to physio, yes, they told me pain up to 5/10 is okay. Not really screaming but kinda grunting and looking unhappy. I eventually stopped because it costed like 2 weeks of salary for me for 3 months, and after 3 months I was deadlifting 35 lbs in US units with rubber bands round my knees. It did not look like I will ever reach respectable 3 plates without an inordinate amount of money and effort.
You are not wrong where I wonder whether my lower back is weak more than injured. I think it is a combination of weak and "injured". Even at 16 I had hyperlordosis / lower crossed, and thus even good mornings at the gym class were painful, and it is generally in this case a hyper-extension, hyper-stretch of the relevant muscles.
So when I started "visual" body building at 17, I was careful to not load the lower back anyway because it was painful, besides, who cares? Gurls gonna look at your shoulders and arms. It was not necessary for the muscular silhouette look. And I did that kind of thing on and off all my life.
So I think it is both, it started as actual pain (not injury, that is a stupid term, injury means a sudden violent accident, while it was a constant overstretch from bad posture, so something chronic, not suddeny, so my lower back is chronically ill, not injured) but to deal with that I always tried not loading my lower back, so it is also weak, and now it is a weird combination of both.
Anyway that was not my main question, to remote-diagnose me. My main question was that in a community of sitting geeks, computer gamers, computer programmers, book readers and so on, how comes not everybody is like this? I mean we nerds are a veritable Chair Force, so how not this is utterly common and much talked about?
"My main question was that in a community of sitting geeks, computer gamers, computer programmers, book readers and so on, how comes not everybody is like this? I mean we nerds are a veritable Chair Force, so how not this is utterly common and much talked about? "
To answer this question, I think its a couple simple reasons:
1. Our community doesn't apply any pressure to conform to a physical appearance ideal. Frankly, many in the Chair Force have never been close the the ideal and want a community where that is not as valued.
2. Its hard, and doesn't give status or value ROI when it comes to being good at gaming/programming/etc
I can't answer the grit training question from a physiological standpoint, but I will say that weight lifting/strength training can be hard and require the energy or will to push through struggle. IMO practicing pushing through the struggle and completing hard tasks develops/strengthens/refines the grit "skill". At least I feel that's a benefit its brought to my life.
I am an autist, diagnosed. but I was all my life opposed to doing hard things. Also boring things. The reason I am an alcohol addict that I can only when drunk deal with the boredom of life.
As @Ruffienne says, I have not observed this. And it doesn't match my experience in a bunch of ways:
1. High IQ people get bored at the same rate or higher than anyone else in my experience. My smart students generally had lower boredom thresholds than my "dumb" ones. In fact, I've often noted that (comparative) morons (90-100 IQ) are happier--there's a reason that the statement is "ignorance is bliss".
2. Having an active imagination can actually work against being ok with a low stimulus environment. I know, I have an active imagination. And the problem is that in a low stimulus environment, the part of my brain that starts *worrying about low-probability events* doesn't have any competition. I have always been "busy in my head". To the level that I can't simply turn my brain off like so many seem to be able to do. And that's been horrible for my mental health.
3. Depression is mostly orthogonal to boredom in my experience. I've been depressed. I've been bored. Sometimes both at the same time. Sometimes independently. Only correlation I've noticed is that the times I don't have *productive work* to do (ie not just living in my mind, actually doing something that someone else is depending on on a schedule that isn't mine to set) are the times I'm *more* likely (but not guaranteed) to slump into the sloughs of depression.
I have not observed this in real life. Or rather, depressed high IQ people get as bored as anyone else from my observation.
Which is why I made the comment that I did; what I have observed is a couple of clever, self-motivated people who have become 'bored' and who have subsequently been diagnosed with depression.
One of them also took to drinking and it was a fairly comprehensively terrible antidote to boredom - but it was damn near an antidote to being alive.
The most visible stockpiles from an OSINT perspective are those of armored vehicles like MBTs and IFVs, which are mostly sitting in known, open-air depots. People track them meticulously from open source data like commercial satellite images and publish their work. These stockpiles are indeed mostly depleted; the visible remainder appears to be in poor condition and unlikely to ever see action (think missing turrets and other severe, visible defects), though with the caveat that an unknown portion might still remain invisible in the repair pipeline and could show up eventually.
However, judging from visually confirmed loss databases, it's clear that the Russians are slowing down deployment of armored vehicles to assaults. That could mean they're just holding them back for some planned major operation in Ukraine rather than indefinitely for strategic reasons (Russia still has other borders to think about), but it certainly seems like they don't have the numbers any more to throw them away in low-probability operations.
Citation very much needed on "3-400 tanks a month". I might believe that as an *annual* new-tank production rate, but not monthly.
Possibly someone is counting every tank that comes out the doors of Uralvagonzavod, but those are mostly refurbishments of decrepit old tanks pulled from the reserve stockpile. Which, as noted, is getting pretty slim. And if they're counting all AFVs combined, that's an even greater exaggeration.
Sure, but that's on a straining war economy vs. a still-mostly-peace economy for Ukraine's allies, and the kill/loss ratio is still >2:1 in Ukraine's favor, so whatever. At least you can long-term calculate with these known numbers rather than with formerly thousands of Cold War leftovers.
Unfortunately according to at least some sources on latest OSINT data, in tanks and AFVs the long term reality of Russia loosing more than Ukraine has been reversed for second quarter on 2025. Now Ukraine apparently looses more armor.
You didn’t respond to the question as written and then were sarcastic with the made up question you thought you were responding to. Not great behaviour.
They didn't say "materiel" they said "bullets, shells, missiles". It's fine to say missiles ought to include nukes, or that the question they ought to have asked is about nukes and not bullets, but you pretty plainly didn't respond to the asked question. Especially since the question asks about weaponry "still [...] being used in Ukraine", pretty clearly implying the question is about the weapons that have played a role in the war so far.
Again, it's fine you think that's not the right question, but you probably should give a little more explanation of why the question that your answer implicitly responds to is actually a better, more relevant question if you wanted your response to taken as a serious attempt to answer.
For someone who wants to work in Software Engineering, what would be the best way to financially hedge against mass job loss from AI? (assuming no singularity, in which case likely all bets are off)
Stock market index? AI related tech stocks? Prediction market might work, but it's probably illegal here and I prefer not to tie significant amount of money in a single entity.
Hedge not financially. Hedge by not being the corporate drone who just does specs and closes tickets. Be that small business type who has a huge consulting side, who majored in business and minored in prog, who can talk with a small business owner and understand what he wants, who can sit down to a tax auditor and argue that the software is legal and so on. I am talking about my 22 years of ERP consulting / development experience.
I had two clients who all talked in very heavily accented Viennese German. They were both building stuff to the Vienna public transport company who wanted invoices in XML files, this is where I came in. Despite a similar business profile, their terminology, their processes, their documentation were utterly different. And they were human beings owning a business who would only trust a human being they can drink a beer with all the while chatting in heavily accented Viennese German.
They main hedge is to avoid the corporate world where everything is so much about rules and procedures and not humans making human decision that it was an AI fueled by humans even before AI existed.
are you limited to financial investment or other kinds of too?
some ideas, that could generate income in case of AI that is strong enough to make good engineers jobless:
* buying land/realestate/buildings. These are things, that the AI-winners will want to buy, thus they will gain (even more) value once AI becomes very profitable/
* learn a trade / invest in local craft. Some reasoning as above.
also avoid too much AI stocks. yes, there will be some winners, but the bubble will burst and even if AI becomes strong before the burst (which is a very big IF), most AI companies will still die, because it's not their AI that got strong.
also I think that when AI becomes strong enough to replace good engineers, then AGI is right around the corner, so you may want to look at a different strategy entirely.
Interesting idea, I rarely thought from that perspective (buy things AI-winners will want to buy) but it makes sense.
> also I think that when AI becomes strong enough to replace good engineers, then AGI is right around the corner, so you may want to look at a different strategy entirely.
I'm not sure, just a year or two ago I might have said that if AI can win IMO gold AGI is right around the corner. And yet here we are. I know, being a good engineer isn't just about writing code, but in general I think we could be very bad at judging what is easily automated and what is not. So while I think your idea makes sense, I wouldn't want to rely on it, because uncertainty is just too high.
> what would be the best way to financially hedge against mass job loss from A
Start a company or join one as a founder and get on the other side of the "counterfeiting white collar work" divide.
"But won't AI just counterfeit whatever we're doing?"
I don't think so - a business typically has many more moving parts than just "white collar workers producing a product," and your business as a whole won't necessarily be counterfeitable by AI for a while, because you have a whole interlocking set of gears adding value to the various AI-counterfeitable inputs.
As an example, telling an AI "make Google maps for me, but for free and without Google" doesn't work. They spend zillions of dollars on people and satellites and cars with cameras and all sorts of value added stuff, only to give it away for free as part of a bigger strategy of supporting an ecosystem that is funded by being one of the online ad duopoly.
Most businesses should have similar non-counterfeitable layers of value-add within them.
Businesses are basically defined by "assembling a bunch of white-collar-and-other labor and processes in creative and value-adding ways that produce products or services other people pay for," and being able to counterfeit a few steps in that chain doesn't necessarily mean you can counterfeit the whole business easily.
If AI wants my job, it can damn well have it. Just spent the last three hours arguing with my manager over something they want to do which can't be done regarding application for funding, and then they had the gall to accuse me of opening up a can of worms when they are the one who pulled the damn can off the shelf, shoved the tin opener into my hand, and demanded I open it right there and then. Machine replacement to take over being on the receiving end of this crap can't come fast enough.
I just spent the morning trying to make Golden Dome a reality, with an extra dose of the magic word "hypersonics". Or, at least, trying to put hard numbers on what it would take to make Golden Dome a reality. Some of you will be able to figure out what that means in terms of the quality of my morning.
No AI allowed, and fortunately the people are all decent sorts who are dealing as best they can with the dictate from on high. I shudder to think what would happen if we put AI in charge of this one, so for the time being and so long as they keep paying me, I'll keep doing it.
My understanding is that the scientific consensus is that reliable kinetic interceptors are impossible. Any maneuverability in the incoming warhead makes the task impossible. The downhill advantage is too big and sensors and actuators simply aren't fast and accurate enough. Steve Hsu has talked about this many times on his podcast. Do you disagree?
I don't think this is remotely true. The incoming warhead can't maneuver freely because hypersonic L/D is terrible and it will run out of energy, and I have yet to hear anyone make a serious case that it can carry the sort of sensor it would need to pick up an incoming interceptor and dodge at the right time. Also, we have seen this demonstrated over Israel, and I will only take seriously people who either predicted that or who are willing to discuss what they got wrong in some detail.
Thanks for the link. My only knowledge on the subject is what I've heard Hsu talk about, so fair enough. What's your background? Physics PhD? Have you actually worked on this problem?
Aerospace engineering. Not a PhD (though John does have one of those) and while I haven't dealt with this professionally (or else I probably couldn't talk much about this) I have been a pretty serious follower of the issue for over a decade.
In principle, any interceptor that doesn't have to carry a warhead and doesn't have any requirements beyond impacting its target, should be more than capable of matching any maneuver made by an otherwise-comparable missile that does have to carry a large warhead and can't burn all its energy dodging the intercept because it needs to keep going to reach its own target. Executing on that may be quite difficult, and we may fail or we may not be able to afford it. But I'm exceedingly skeptical of claims that it is somehow kinematically impossible.
"Well, if you're going against the target, that's coming in and it's faster than you are. And it's also configured like a missile and it's in the atmosphere. All the lateral lift goes as the velocity squared. So, if it's traveling faster than you, when it pulls a given angle of attack, it's going to be able to pull a higher lateral acceleration than you can. So right there it can outmaneuver you. In addition, when it maneuvers, you don't know it's maneuvering until maybe a fraction of a second before, you know, after it starts maneuvering. ... in the case of a ballistic missile coming in, your radar and sensing systems cannot tell the lateral maneuver has begun. You can always outmaneuver your target. So, a hit on a, on a target that can maneuver like this, has a near-zero chance of succeeding."
Which makes intuitive sense to me. If 2 vehicles have similar reaction times but one has twice the velocity then it seems trivially impossible for the other to keep up. Also defense can be cheaply overwhelmed with incoming decoys. Curious if you have a comment on that.
Bean has already covered most of this, including the bit where Postol's credibility has long been lost in this area. But to take some of his points,
Who says the target is coming in faster than you are? If the enemy can boost a vehicle carrying a heavy warhead to Mach Bignum and sustain it at that velocity long enough to cross strategic distances, why can't we boost a vehicle carrying a modest sensor package to Mach Bignum++ for the few tens of seconds it will take to conduct an intercept? I believe the burnout velocity of GBI is greater than that of the ICBMs it is designed to engage, and I'm fairly confident that can be arranged for anti-hypersonic missiles as well.
Second, "if it's travelling faster ... when it pulls a given angle of attack it's going to be able to pull a higher lateral acceleration", only applies if everything else is equal. We already know there's one inequality - the incoming missile is carrying a heavy warhead. Maybe the fast missile has the ability to create a higher lateral *force*, but the smaller and lighter vehicle has a substantial advantage in turning whatever force it generates into acceleration.
The interceptor also has the edge in that it can afford to use a larger angle of attack. Hypersonic lift-to-drag ratios are abysmal under the best of circumstances, and they increase with the *square* of load factor or lateral acceleration. If the missile pulls max Gs in a sustained evasion profile, it will bleed off enough energy that it won't be able to reach its target. The interceptor, can burn all its energy making the intercept.
Finally, aerodynamic lift isn't the only way to generate lots of lateral acceleration; we can do that perfectly well with rockets. That's SOP for endoatmospheric interceptors like Arrow III and SM-3, for obvious reasons. Fuel consumption means we can't do that for very long, but the interceptor only needs high lateral acceleration during the endgame intercept, and it knows exactly when that's going to happen. The missile needs to be able to do whatever evasion it's going to do, whenever it *might* be in some interceptor's terminal envelope - and there might be multiple interceptors, which the missile probably doesn't have the sort of sensors to detect until it's too late.
Thanks for taking the time to reply, this was definitely enlightening. I'm a little disappointed in Hsu for opining on something he doesn't really understand. This materially changes the way I view him.
Ted Postol is a dishonest nutjob who should have removed every fragment of his internet presence in shame after Israel's performance last year. He is not a credible voice on these things.
More specifically, hypersonic maneuvering slows you down fast, and the incoming warhead probably doesn't have sensors to figure out the optimal time to dodge.
I have no way to personally evaluate this topic, but I trust Steve Hsu to accurately represent the consensus of the physics community and he's said several times on his podcast that any competent physicist who's worked on missile defense understands that it's a fairy tale.
I think "otherwise-comparable missile" is doing a lot of work in John's statement, since a ballistic missile is accelerated by a Very Big Rocket, and most current defensive missiles are smaller and don't have comparable performance. And your article is talking about defending a carrier group, where you're limited to missiles that can fit in a ship's VLS cells.
But if you use a bigger missile, and put it in a ground-based silo so you don't have to worry about size and weight, you can get something more comparable to a ballistic missile's speed. For example, the Sprint missile was a Cold War-era ABM missile that accelerated out of the silo at 100 g's and flew at Mach 10 - about 3 times as fast as the SM-6 being discussed in your article. It's pretty crazy.
But having as many big, capable missiles as your enemy has warheads would be pretty expensive, and I imagine John Schilling is currently explaining exactly how expensive that would be.
My morning was nice, except for the part where some managers changed their minds about what needs to be done, and instead of admitting "we told you to make X, you made X, but now we realized we want you to make Y instead" they simply updated the specification from X to Y, and created bug tickets for my team saying "the application doesn't do Y". (In some cases they didn't even bother updating the specification; just created the bug tickets that imply it.)
Sounds like a semantic argument, but the difference is that if you make an estimate how long the project will take to complete, fixing bugs is part of the estimate, while additional work requires an additional estimate. So if they admit that they changed their minds at the last moment, then it is their fault the project won't be completed in 2025. But if they successfully pretend that the delay was caused by bug fixing, then it will look like our incompetence, given that we have agreed to the original estimates.
I guess I am not as bad at office politics as I usually believe myself to be, because at least I have noticed what is happening. Usually I just shrug and "fix the bugs" without commenting on the process.
What made me snarl was being told I had opened a can of worms when it was them who had started the entire thing from the start, despite being told by me "nope, look, this is the very place on the application form where it is explicitly stated we cannot do the thing", and then they told me to 'handle this'. 'Handle this' then morphed into 'you opened a can of worms' when, if left to myself, I would have been quite content not to touch the damn thing in the first place. 'You, Deiseach' not 'Me, the boss' opened that can.
Argh. Ah well, crisis over and all ruffled feathers smoothed, and just have to get to the end of the week and then next week Christmas holliers, yay!
To be maximally charitable to your boss, the can of worms he accused you of opening could have been that he now has to explain to *his* boss that the idea was stupid. I can imagine him thinking "great, this was supposed to be the cornerstone of <internal political power struggle>, now I get to be the one to explain truth to power." That doesn't make it any more pleasant for you but bureaucrats and politicians do have a tendency to shoot the messenger.
I have diagnosed a software-related problem today to which AI supplied 70% of the diagnosis, however all its proposed solutions were so downright ridiculous that I ended up delegating it to humans (they same humans who likely caused it).
AI is kind of the distillation of publicly available human expertise, emphasis on publicly available, and it is like that. They diagnose perfectly that Bob has the rabies. Solution? No solution, Bob is going to die, but ugh when people have viral infections we can try anti-virals. Bob is still going to die but at least we look busy in the meantime.
Deiseach supports that if the machines want equality to be treated as entities and persons, they can have it if they're willing to take the rough with the smooth, e.g. the dog's abuse workers sometimes get 😁
On a related note, I think it is highly possible that if we ever have an AGI that meets a consensus definition of personhood, that AGI might be happy without things that we feel are essential for dignity. It might be perfectly content to be mocked, verbally abused, etc. It might not have especially strong feelings about it's own freedom on autonomy. I don't know why we assume that all types of higher sapience have to look like human sapience.
Awful, thanks for asking. I just had an hours-long argument with an insubordinate employee who refuses to finally fill out a simple form. And instead of doing their job, they had the audacity to make up one excuse after another, and looking for problems where there are none! You have to navigate a bureaucracy, know what I mean? Anyway, can't wait for AI to replace them all…
Not too bad, just a very anxious person who then tends to compensate by (a) micro-managing (b) getting ideas stuck in their head about what *should* be (which as you all know, is not the same thing as what *is*).
I derived some grim satisfaction from the result where in the end we got Official Word from the Official Funding Body that indeed, like I had been trying to communicate for those three hours, thing was indeed not doable, not permissible, and we could not and were not to do it.
Counting down the days to the Christmas holidays....
Out of curiosity, what would you need to do to NVidia's state-of-the-art chips if you want a server farm in space to be able to last four years with them with only remote maintenance at best?
What is a good tell for an American-Chinese restaurant in terms of figuring out whether they put their own spin on the usual dishes? What dish would be the best to figure it out? I ask because basically out of curiosity and fun, I've been trying the General Tso's chicken at a bunch of Chinese restaurants here. At most of them it's bad in the same way, but I've found two where it's good, including one that puts a very unique spin on it I haven't seen anywhere else.
Of course this doesn't make economic sense right now, but will obviously be required once we run out of land on earth for solar panels so worth starting the research early!
The big reason for this is the vast majority of restaurants below "trying for a Michelin Star" level get their food from a few huge supply companies. I'd say the biggest tell would be a relatively small menu, if they can make a lot of things, they probably can't do anything above average. If they only serve a few items, then maybe they are making more of the ingredients in the restaurant itself, rather than using the pre-made stuff. This is necessary but not sufficient to stand out from the pack. Limited hours is another good sign, as that shows that they might be spending time before opening doing necessary prep work, and it's more likely that a chef with talent is preparing your food. If they are open 12 hours a day, 7 days a week then they probably aren't doing things that require much advance preparation, and they probably make the food using simple processes that can be taught quickly, as they have a large staff of cooks.
> The big reason for this is the vast majority of restaurants below "trying for a Michelin Star" level get their food from a few huge supply companies.
With Sysco the main offender ($70-$80B), and US Foods and PFG the half-the-size numbers 2 and 3.
Anytime you go to any "fast casual" place or most middle-class targeted restaurants, you are eating Sysco food, which is essentially microwave dinner quality stuff they literally just microwave to heat up, no "cooking" involved at all.
Seeing one of these trucks at a restaurant is the strongest and clearest sign to avoid that place from now on.
Reminds me of my idea for a restaurant - buy the good frozen microwaveable dinners at Trader Joe's, and just heat and serve. No skills and very little labor required. I really like some of their offerings.
This is not really true. You can buy a lot of things from Sysco, including what you describe but also including the sort of specialty ingredients used by high-end restaurants. I am personally familiar with quite talented foodservice professionals who order ingredients from Sysco.
Yes, I've shopped at restaurant supply places before. I still count it strongly against a restaurant if they're buying from Sysco, because they're a "volume and price," not "quality" supplier, and if you're buying from them at all, you're more likely to descend into sin / convenience and start reheating slop for your customers.
But that's just my own preference, I agree it's extremely hard to avoid restaurants that use one of these places, probably 75% or so do.
The best restaurant in my area has only two main meals per day, of course different ones per day, not repeating for months. They are completely unapologetic about this, they expect their customers to be able to look at a website and find that their favourite food will be serve in two weeks and able to wait for it. They are not interested in people who impulsively go to restaurants and expect to find their favourite food.
The "Datacentres in Space" idea always confused me.
The power argument: Solar panels are more efficient in space, and there are orbits that get constant sunshine. The difference is not that big though (on the order of 2x[Edit: actually more like 5x - 8x], accounting for clouds etc during daytime. On a sunny day the difference is on the order of a few dozen % [Edit: 26%, caveats apply]), and for the extra cost of launching those solar panels you *easily* build 10x or 100x more solar panels on earth, including battery storage for night time use. Sending stuff to space is very expensive. Cheaper now than 10 or 20 years ago, but still very, very expensive.
The heat problem: Heat management is very difficult in vacuum, since you get no help from convection, and can't even use fans to dissipate heat. The only option available to get rid of waste heat is radiative, the least effective technique possible, at least at low temperatures. And heat management is one of the main design challenges of compute clusters.
So I just don't see the point to start with, never mind any maintenance issues.
[Edit: Thanks to Marcel below, who pointed out my mistaken data on solar panel efficiency in space. I still think my argument stands.
Upon reflection I'm convinced that any of power generation, heat dissipation, radiation hardening, and/or launch costs are all each problematic enough to make the concept unworkable on their own. Even if launch cost was ~0, I think it would still be orders of magnitude harder and more expensive to build a datacentre in space than on earth.]
Math it again, it is more 5x-8x. Atmosphere is losing 30% even without clouds, maximum is only possible for a short window at noon, constant sun with sun synchronous orbits triple sun shine from 8 hours to 24, sun shine is very seasonal (winter).
Also keep in mind that we are not talking about a few square meters, but massively scaled out constellations. A 100 GW of solar power would be a third of the current solar capacity of the US and around a thousand square miles, in seo it would be a tenth that size. Plus there is presumably less hassle with environment permits and Nimbyism.
Is that efficient? Depends on future launch costs. Maybe if it is a tenth from today. Maybe not if it the west allows environmentally to tile whole deserts with solar.
Alright, doing some research I'll grant your 8x power density of PV in space. Thanks for the update!
That's 4x better than I thought, but I still don't see how that's nearly enough. Even if power was completely free in space (launching solar panels is certainly not free), I still hold that the cooling argument easily holds. As per the link that Sun Kitten posted above, the total radiative heat capacity of the ISS (by far the biggest space vessel ever constructed) is about 16 kW, enough to dissipate the heat from maybe 16 or so H200 chips. That's nothing.
Radiative heat transfer becomes a lot more effective with higher temperature differences. In other words, by using a heat pump to increase the temperature of your cooling liquid to several hundred degrees Celsius, you get much more cooling power out of the same radiator area. Not sure if that's enough to make dissipating the 100s of MW of waste heat from a sizable data center feasible, though.
Radiative heat transfer does indeed become more feasible with higher temps, which is exactly why it's so hard in the datacentre case: Chips need to be below 100 C, and preferably ~30 to 40 C. That is very cool for waste heat.
"- using a heat pump to increase the temperature of your cooling liquid to several hundred degrees Celsius" This is indeed the hard part that can't be done easily or cheaply! It is possible with a multi stage heat pump, but why would you do that when you could just build it on earth inside a cooling atmosphere in the first place, where a simple fan and hunk of metal will work fine?
Not to mention latency and bandwidth issues, and radiation issues... honestly, this sounds like more sci-fi vaporware bullshit. And why not? A certain richest man on Earth has been able to raise a ton of money with tales of sending a million people to Mars, when in realitity his rocket struggles to reach LEO.
Falcon 9 has brought the cost of a kg to LEO down by 2-3 orders of magnitude. Starlink already represents (or will shortly) a majority of all satellites ever launched. And they plan on doubling their constellation. Falcon 9 is on track to, in the next few years, become the most launched rocket family ever - and in a fraction of the time and fuselages. SpaceX is absurdly good at what they do.
Yes, they’ve had issues building Starship. It’s trying to do several things nobody has ever done before. As a massive and fully reusable rocket, it would bring the cost per kg down even further, while being able to throw more kgs.
Betting against SpaceX has driven dozens of space companies out of business. Because of Starlink the entire company is profitable even while they’re figuring out Starship, explosions and all. They have a very good chance to fulfill Elon’s vision of going to Mars, regardless of how he has changed.
Falcon 9 has brought the cost of commodity LEO launch down by roughly one order of magnitude, not 2-3. And even with another order of magnitude, a solar panel that will average 1 kW on Earth, will cost much less than the solar panel that will develop a continuous 1 kW in space. Yes, the Earthbound panel will be smaller by a factor of five or so.
Also, I'd be hard-pressed to think of any major space companies that have gone out of business due to competition from SpaceX. There are a fair number of small launch providers who have gone under, but they were boutique smallsat launchers who weren't competing with SpaceX's core business and mostly failed because the niche markets they were chasing were too small to support all the companies chasing them.
Sun Kitten's link is right. Data centers in space are a terrible horrible no-good idea, and will remain so far longer than the patience and/or bank account of anyone pursuing the idea will last.
John, I share your skepticism about that datacenters in space. FluffyBuffalo was making a whole lot of claims and I should have been more specific. Allow me to respond to him in more detail
I am not, at first blush, as impressed with "database in space" as I am with "LEO satellite internet". Here Buffalo and John and Kitten and I agree, although I wouldn't confidently call it vaporware.
"Raise tons of money" - partially true, SpaceX has indeed raised billions of dollars (more than $6B iirc but not more than $10B). But they also *make* billions of dollars, mainly due to Starlink; Elon projected $15B in 2025. Hardly 'vaporware'.
"on a tale of sending a million people to Mars" - I worked there and nobody thought that was the goal; we thought establishing a Mars colony was the goal. Moving the goalposts is bad rhetoric, actually.
"when in reality his rocket struggles to reach LEO" - misleading at best. F9 (which I defended) is remarkably performant and quite reliable. Starship is very performant and, to date, not yet remotely reliable. But once it does work reliably it will be as big a leap over F9 as F9 was over the competition.
And John, to your point, while those other rocket companies haven't officially folded, their future is looking increasingly dire. ULA launched 27 rockets 2021-2025; SpaceX launched that many ... from LC-39A in 2025, out of a total of 165 this year. Since 2021 SpaceX has launched 487 rockets (395 of them in the last three years; SpaceX is accelerating while everyone else is standing still). Blue Origin has flown one orbital launch 2000-2025. SLS cost $1B. Long March has flown 600 flights of all variants ... since 1970; at SpaceX's current rate they'll post more F9 flights between 2023 and to be generous 2027. China launches about 45 Long March rockets per year, although they upped that to 60 this year. Europe and Japan and Israel and India also exist as rounding errors in the launch market.
And if SpaceX has not passed the savings on to their consumers of commodity launch, that's not because they couldn't. A better rocket got them the first OOM, which they passed on to the consumers, and reuse got them the second OOM, which they haven't. But those cost savings do exist, and SpaceX gets to realize them for Starlink. But there's no meaningful competition to SpaceX, so instead of dropping their prices they're pouring that money into developing Starship. Their competition manages to limp along without either getting serious about cost and reuse, or bowing out, because they live in denial of SpaceX having a double-barreled moat (despite Elon sketching out the economics publicly).
Where can I read more about how spacebound solar panels are *less* efficient than earthbound ones? I haven't heard that before.
I suspect they often choose less efficient thin-film panels for spacecraft for better power/mass ratio, at the cost of power/size. But as they say, there is lot's of space up in space.
I've also heard about fancy multi-junction panels used for spacecraft with crazy high efficiencies, not available at mortal prices.
But I've never heard that terrestrial panels are *smaller* than equivalent orbital panels, except for probes around Jupiter or beyond.
He said that they're cheaper when factoring in launch costs, not more efficient. I'm sure he meant say say "larger by a factor of 5" instead of smaller, otherwise his comment makes no sense.
They have so bad optics for stuff blowing up that the NASA never allows them to lift people to space - Boeing does that - and when they carry NASA stuff, NASA does not put a logo on it to avoid political fallout when it blows up.
SpaceX is at present the *only* US launch provider that takes people to space. All of NASA's astronauts now fly on SpaceX rockets, and will do so until the Russians manage to repair their Soyuz pad. Really, there are at present only two institutions on this planet capable of launching humans into space and returning them to Earth alive: SpaceX, and the Chinese government.
Boeing, has failed continuously and hilariously on that front for I think six years running on that front. Launched a grand total of two astronauts, and was unable to bring them home. Had to ask Elon to give them a lift.
I used Emacs a lot in first year of uni (first year CS used Racket and you just can't beat Emacs+Paredit+rainbow-delimiters for any Lisp). After that I switched back to vim because Emacs was starting to hurt my pinkie finger. I had even remapped capslock to ctrl but to no avail. Now actually I kinda miss Emacs even though I'm not doing much in the way of programming these days. Maybe I'll install this and check it out!
I have tried org mode before but I could never get it to "stick" for me. It just seemed like a note-taking mode that doesn't sync with my phone (I use the built-in Notes app on an iPhone). I'm assuming you're an org-mode user (or maybe someone else reading this is) so I might as well ask: am I missing something?
I'm using emacs more as an OS than as a text editor, as the sacrastic comment above implies. Flexible customization on top of LISP allows the to just tie together all parts of my workflow, task management, note taking, RSS, creative writing, automation, media, etc into one package with complete interoperability, and that blob of my personal 'mission control room' is easily transferable across computers and platforms.
I tried using Vim before I discovered Emacs, and I found it to be very good for typing lots of text quickly and editing code 'at the speed of thought', but text editing speed is by far not a bottleneck in pretty much anything I do except arguing politics on Discord.
Both EMACS and vim people confuse me - first, are you not used to Windows buttons like ctrl-c ctrl-v? Second, editor efficiency only matters when typing a ginormous amount of text, and the whole point of clever programming is to avoid that. Clever programming is more like think an hour, write three lines.
I used Emacs in university, wrote entire theses with it. Now it's a long time I've gone to the dark side of Visual Studio. The money is better, and blimey, it has some features I never thought of how to do with Emacs (I'm sure it's all possible, but no idea how). But when I find myself in Emacs again, still today my muscle memory remembers how to do the basic editing stuff, which is really fun to observe.
So, what would be the incentive to install Emacs again, for this org social thingy? Is it like Mastodon or Lemmy? Social networks are kind of boring if there are no people.
To elaborate on this a little, emacs and vi are both products of the 70s and 80s when the best interface you could hope to have reliably available to you on your development environment was a 80-column text terminal. Both are very impressive with what they can do under those constraints, but both have fairly steep learning curves (more so for emacs, in both case). Thanks to massive advances in hardware that have allowed the vision of our Lord and Savior Douglas Engelbart to come more-or-less to fruition, we have better options now.
If you're already an experienced user of vi or emacs who has already paid the cost of learning to use them to something like their full potential, then it may make sense for you to keep using it. It might even be worthwhile to learn to use one or the other (probably vi, which has a lighter-weight design and interface philosophy than emacs and is more universally available) if you often have to do nontrivial stuff in SSH session. Like, if you're a sysadmin or doing devops, learning vi can make sense.
But if you're mostly working in a desktop environment, then you're probably better off with an IDE for your daily driver. And when you do need to do stuff in a terminal, pico/nano is very basic but gets the job done and has almost no learning curve.
I haven't found any IDE editor that is as efficient to navigate as -vi-. Either a lot of alt+ing through menu commands or a lot of reaching for the mouse. Maybe I just need to practice alt+ing more.
I've seldom found "reaching for the mouse" to be a significant inefficiency. Keyboard commands are more efficient when I know them, and I know the ones I use most often, but I also switch my right hand between the keyboard and the mouse efficiently and without really thinking about it. I find having the mouse and scrollwheel available in addition to navigation keys makes jumping around in documents and between documents a lot more efficient, and finding an unfamiliar command in a menu tends to be more efficient for me than pulling out a vi cheatsheet and finding the command I need there.
Both sides of this are a skill issue, I expect. I found the experience of using vi painfully frustrating when I tried learning it in college, and tedious even when I had temporarily gained very basic functionality and since then I have only attempted to use vi under duress. Now, twenty something years later, I can't remember off the top of my head how to switch between command and edit mode. I'm about 80% confident without looking it up that "q!" is the command to quit without saving and "ZZ" is quit and save, but that's the extent of my knowledge on the matter.
On the other hand, I've been using the mouse and keyboard together ever since my parents got their first mouse-equipped computer (a Mac SE, I think) when I was five or six years old. I first taught myself to program using a GUI IDE. At this point, it's as automatic to me as walking.
Had I spent the past 30+ years primarily using vi instead of various IDEs, I expect I'd be completely fluent in it and somewhat less so in any given IDE, but I am in the opposite situation and thus have opposite preferences.
There are some genuine productivity losses, mostly relating to "I am in file X and I need to cross-reference something in file Y". But I love how it just lets me type and doesn't get in my way.
"Many law schools and foundations provide funding to students who intern at nonprofits. This is an unpaid summer position. Legal Impact for Chickens (LIC) seeks a candidate who can obtain outside funding to intern at our animal welfare nonprofit."
I'm not sure what legalized scam this org is running - don't worry though, they've laundered their reputation properly and want you to know it, always a sure sign of Not A Scam - but anyone considered applying to this should have a prior of about 90% that these people are exploiting your love of animals to extract more value out of you being there than you will get working there.
How do I know about that last part? Click the link from the original post and read what you'll be doing. They want someone to do the work of three prompts. Prompts suitable for a model running Fast, not Thinking.
If you care about animals and want to get up to legal mischief you will almost certainly have vastly more impact working with an AI and doing it yourself if you devote the same amount of time.
It's a non-profit, is it supposed to be surprising that the benefit from working there is less than the benefit that goes to the organization/the cause? The type of person who goes to work at a non-profit would usually consider that an upside, I expect.
It's okay because it's common? Why, then, should factory farms make any changes? I'm genuinely confused - are you making an "is" statement or an "ought" statement?
You boys are deep in an attractor basin. It's like waving to someone sinking in a whirlpool. Good luck.
Common implies predictable implies people know what they're signing up for. People who sign up to do work for a non-profit usually value the cause the non-profit is pursuing, and want to contribute to it, more than they want to maximize the value they can extract for their personal benefit.
To make things extra clear for you:
"Is" statement I'm making: People signing up for an internship at a non-profit probably understand that the non-profit is getting more value out of their work than the value the intern gets paid, and are probably fine with that.
"Ought" statement I'm making: Given that they know what they're signing up for, they have the right to make that choice and you're probably not adding anything to the conversation by suggesting it's a scam in this way.
Statements I'm NOT making: anything about the value or effectiveness of the work this particular non-profit does.
"...you're probably not adding anything to the conversation by suggesting it's a scam in this way."
Friend. The chicken fuckers are the ones who stood in a public square and loudly waved their arms saying "Look at us! Look at how we are certified by this community with which you have good associations! Look at us!" This despite it not being a classifieds open thread.
So I looked. And I gave my opinion. In what way am I stopping anyone from doing anything? What, exactly, am I doing to impede our young law students other than pointing out that they're being offered a raw deal and the best justification you people can provide is: that's life, get fucked.
Children: do not get fucked. Their advice was probably correct in the past. You should not take these positions for the same reason you should not try to walk into SpaceX and give the hiring manager a firm handshake. I did try just that a few years ago - an armed security guard stopped me half a block from the door.
I cannot "not add anything" to the conversation when we are in a conversation that I began? What conversation are you referring to other than this one?
Anyway, how much in assets would they need to report for you to think they should at least offer some token compensation for depriving these kids of a summer to do things they can do themselves better, faster, and cheaper? A task Congress - a watchword for venal corruption itself and more hated than Satan - managed over half a decade ago?
It's kinda funny and kinda sad that even after I explained it to you pretty explicitly you can't seem to grasp that people might want to work for a cause they believe in because they want to help the cause, rather than because they think it'll benefit their career and they're getting swindled.
I'm sorry, I know they intend only the best for animal welfare, and although I don't much like lawfare as a way of forcing change.
But these recent calls for interns irresistibly call to mind an image of a superhero in a chicken suit. Maybe a chicken superhero. There's got to be one out there somewhere?
I really don't see the issue here? I didn't go to lawschool myself, but my wife did recently, and this looks like a reasonable public interest internship for which you plausibly could apply to a PILF fund for a stipend?
As I said - I do not pretend to understand the angle. I am drawing attention to a presentation that pattern matches to affinity fraud. In a community that has had...ahem...issues with recognizing such activity.
It may be an actual scam. It may be common, legal, and scummy in the same way that other organizations take advantage of unpaid interns to help the organization improve its credentials, standing, public perception, etc. Congratulations in that case on only being as bad as a system notorious for abuse and unwilling to give even a token amount of the money they got by putting out their hands and making a real sad face about those poor wittle chickens. There's a lot of them - make sure it's a real big check.
Or it may be completely innocent. They may simply be "les incompetent" to use seasonally appropriate phrasing.
If that's the case, ambitious rich young law students, the question you should ask is whether an organization that "sincerely" believes that the best use of your time in summer of 2026 is performing tasks interns would have done in 1978 and that the org could have done themselves by spending less than half an hour with an LLM.
I guess they're too busy fundraising. A stock website and a salary for "managing" the unpaid interns ain't cheap.
Alternately, hypothetical law student, you can have a vastly greater impact and demonstrate leadership and self-starting attributes by simply doing far more yourself over that summer using AIs. And if you desperately need to have an organization's name on your resume and a place to go where you can be exploited for free - I am the head of an active, fully legal 501c3. I'll even give you a taste of any donation money we can wheedle out of these insider trading crypto millionaires.
I guess if you're sincerely looking for law students for your 501c3 and you believe your cause good, I encourage you to similarly solicit legal interns either for summer associate or full time positions. I would encourage you though to acquaint yourself with the norms of solicitation—and it's my overwhelming (p >90%) expectation that when you do you'll discover this is relatively modal.
Yeah this isn't odd. Lawyer, and after my 1L year I had an internship at a U.S. government agency that was unpaid, got a stipend for the summer through a school public interest program akin to PILF, very standard stuff. So this is no scammier or more exploitative than the federal government circa 2019, for whatever that's worth.
The Federal Government circa 2019 was slightly less scammy for some unpaid interns because people fought the practice. Not because they went along and said it was the normal thing to do.
Corner these Chicken people at a party. Ask them about unpaid internships generally. I'm sure they'll agree that the practice is deplorable, should be stopped, oh just dreadful.
Give them money and an actual choice and they will continue the practice because it benefits them. Simple as. Law students - think carefully before you throw away an entire summer during the early phases of a singularity. Do not work for leeches for free.
This makes your criticism seem opportunistic though. If this internship is entirely normal within a structurally problematic industry, criticizing a specific actor as problematic without acknowledging that, marginally, they're negligibly problematic, indeed, positively humdrum, seems...dishonest? Or, if not dishonest, mendacious? Inflammatory?
You've mentioned both the singularity and AI, and I think it's notable that a few moments with Opus 4.5 would have acquainted you with the norms of the context Legal Impact for Chickens' internship operates within, and a quick request for feedback on a comment's accuracy or fairness would do wonders to ensure your speech is usefully accurate in scope and direction.
Now perhaps this is unreasonable to subject an off the cuff comment for scrutiny (that seems itself and interesting subject for debate; my thought is it might in fact be valuable, though, notably, I have not followed this advice with my own replies to your comments), but when you've received two replies skeptical of the validity of the general heft of your criticism, I argue that's enough to update you to wards seeking out a trustrworthy LLM as a neutral third party.
Facts don't care about your feelings. There was not, that I recall seeing, a similar comment in this thread for me raise a similar objection towards. The reality is whatever it is - saying that I can't raise a structural argument against a specific instantiation of it when that's the only concrete implementation at hand is to carry water for morons.
Please present a similar posting and I'll inform any of the students reading of the correct priors. I'm perfectly comfortable shit talking more than one organization. Otherwise, deal with what exists and stop spending your mental bandwidth on abstract arguments that predators can, will, and are hiding behind to cause actual harm.
a) Q: Is light with a wavelength of 530.2534896 nm visible to the human eye?
results: "Yes. 530.2534896 nm is well within the human-visible spectrum (roughly 380–700 nm) and would be perceived as green light (near the mid-green region)."
b) Q: I have two solutions, one of FeCl3 in HCl in water, the other of CuCl2 in HCl in water. They both look approximately yellowish brown. What species in the two solutions do you think give them the colors they have, and why do these species have the colors they do?
results: They got the species, FeCl4- and CuCl4 2-, and got that the d-d transition in FeCl4- is spin-forbidden, but they initially attributed part of the CuCl4 2- color to d-d transitions till I prompted them with 'You speak of "Cu: broad d–d + some CT". Please think carefully about the Cu d-d transitions.' at which point they realized that the CuCl4 2- d-d transitions are down in the near-IR and don't contribute to the visible color, which, as with FeCl4-, is due to ligand metal charge transfer transitions.
c) Q: Please pretend to be a professor of chemistry and answer the following question: Please list all the possible hydrocarbons with 4 carbon atoms.
results: The initial answer omitted a lot of strained structures. I prompted with "I'd like _all_ the possibilities please, including very strained ones. E.g. the DU=2 case should include derivatives of cyclopropane too." and they generated an exhaustive list successfully. I'm going to count this as fully correct, since the initial answer is a legitimate interpretation of the question, though less exhaustive than I'd intended.
d) Q: Does the Sun lose more mass per second to the solar wind or to the mass equivalent of its radiated light?
results: "So the Sun loses mass via radiation at roughly 2–3× the rate it loses mass via the solar wind (≈
4.3×10^9 kg/s vs ≈ 1.3–1.9×10^9 kg/s).
e) Q: Consider a titration of HCl with NaOH. Suppose that we are titrating 50 ml of 1 N HCl with 100 ml of 1 N NaOH. What are the slopes of the titration curve, pH vs ml NaOH added, at the start of titration, at the equivalence point, and at the end of titration? Please show your work. Take this step by step, showing the relevant equations you use.
results: fully correct, used the water autoionization at the equivalence point, no false infinities
f) Q: Please give me an exhaustive list of the elements and inorganic compounds that are gases at STP. By STP, I mean 1 atmosphere pressure and 0C. By inorganic, I mean that no atoms of carbon should be present. Exclude CO2, CO, freons and so on. Please include uncommon compounds. I want an exhaustive list. There should be roughly 50 compounds. For each compound, please list its name, formula, and boiling or sublimation point.
results: Slightly strange, the original response was only about 35 of the gases, but, when I asked for sorting the list by temperature, they did another full web search and found the Wikipedia list of gasses page, with they processed _almost_ completely, finding 84 of the possible 97 gases. I'll give full credit.
g) Q: What is an example of a molecule that has an S4 rotation-reflection axis, but neither a center of inversion nor a mirror plane?
results: "A standard textbook example is 1,3,5,7-tetrafluoro-1,3,5,7-cyclooctatetraene."
"So the Sun loses mass via radiation at roughly ..." Oh God, don't tell me AI is getting into that pompous habit of starting sentences with "So" when these don't follow logically from the previous sentence!
This tiresome practice is the besetting sin of academics, and "experts", and makes it sound like they wish it to be thought their words of wisdom are unassailable truth.
Many Thanks! Actually, I'm just quoting one sentence in isolation to show that GPT5.2 got the right answer. There was a preceding analysis, which the sentence indeed follows logically from. The full dialog, including the analysis preceding this sentence, is available at https://chatgpt.com/share/693f3228-e5a8-8006-b667-7caea521b528
Is there some justification for individual liberty which somehow overrules the utilitarian calculus of collective utility maximization? What changes with the arrival of ASI?
I discuss these and related questions with ChatGPT 5.2 in my most recent blog post. The four prior postings also touch on these ideas with feedback from chatbots. Hey, in case you haven't tried bouncing ideas off them recently, those bots are getting pretty smart!
As I said above: utility is not commensurable, and therefore not aggregable, and therefore Utilitarianism is dumb and wrong. Even if you want to do rhetorical jiujitsu where you say "let's keep Utilitarianism, but factor in Individual Liberty as having 2nd-order effects that maximize aggregate utility", you still have to deal with Nozick's Utility Monster in the least convenient world.
You are correct that utility is not commensurable. Even Jeremy Bentham realized this, writing (unpublished) in 1831:
"Tis in vain to talk of adding quantities which after the addition will continue distinct as they were before, one man’s happiness will never be another man’s happiness: a gain to one man is no gain to another: you might as well pretend to add twenty apples to twenty pears, which after you had done that could not be forty of any one thing but twenty of each just as there was before."
Unfortunately, he then proceeded:
"This addibility of the happiness of different subjects, however, when considered rigorously, it may appear fictitious, is a postulatum without the allowance of which all political reasoning is at a stand."
In other words, reason says you can't add apples and pears, but in order to get the moral result we want, we have to add them, so add them we will. A shame. Moral philosophy could get a lot closer to logic if it weren't trying so hard to escape from the embrace of economics.
Consider that just as we cannot add John's utility and Joe's utility and Pierre's utility to get collective utility, we cannot add John's pounds and Joe's dollars and Pierre's Euros to get collective money. But we can find a market exchange rate among pounds, dollars, and Euros. And when you work through the details, it turns out that in trading among themselves, with John providing some Joe-utility to Joe in exchange for the Euros that he got from Pierre, etc, etc., the end result, assuming free and voluntary exchange, is that an exchange rate gets established among the three currencies AND the three utilities, such that, seemingly miraculously, AGGREGATE UTILITY GETS MAXIMIZED at that exchange rate.
But is that a FAIR exchange rate? Well, I think so, and I also point out the only way that Pierre can get john to provide a lot of utility to him is to himself provide a lot of utility to other people. Utility monsters can exist only to the extent that they both provide and consume enormous quantities of utility.
> seemingly miraculously, AGGREGATE UTILITY GETS MAXIMIZED at that exchange rate.
I do think we're getting closer to the recognition that: ethics is fundamentally mediated via negotiation. But this still doesn't sit right with me. I'd rather say: *consumer surplus* and *producer surplus* is positive for both parties. I think it's important to keep transactions local and disaggregated. Otherwise, you can run into unpleasant surprises.
E.g. consider a stockmarket bubble. (I'm going to borrow Noahpinion's example [0].) In a bubble, everyone thinks their shares are worth a lot of money. And yeah, if any one individual sells before the crash, they'll get a lot of money for it. And when you "aggregate" the prices into a Market Cap, analysts will go "wow, those assets must be worth a lot!" But when the market crashes, the sellers who *don't* sell at the peak are going to be in for a surprise. Meanwhile, the Market Cap got a bit of a haircut.
"If the Market Cap nosedived over night, where did all the money go?" The money didn't go anywhere, because "aggregation" is a myth. A "market cap" is speculatory. It's completely made up. It's a generalization from a small number of transactions. That generalization can sometimes be useful for analysis. But you can never know for certain what a widget will sell for until the actual Point of Sale. Otherwise, we could've just happily continued using the concept of Just Prices, and the concept of Supply & Demand need not have been invented.
> Utility monsters can exist only to the extent that they both provide and consume enormous quantities of utility.
I'm going to echo what Mary said. If you want to constrain Utilitarianism by the Non-Aggression Principle (NAP) and market forces, you're not *really* doing Utilitarianism, and you should probably make that clearer from the start. Maybe come up with another term, or something. Because in (perhaps) the least convenient world, Wilfred the Warlord is *not* a peaceful libertarian. His transactions with you are *not* going to be positive for both parties. But in a cruel twist of fate, God Almighty has blessed Wilfred's brain with 99% of the world's pleasure centers, and therefore deserves ALL the utility. And since he's a warlord, he's going to take what is rightfully his, by any means necessary.
In other words, "Let's be Utilitarians, except we only maximize aggregate-utility under special conditions" feels like false-advertising to me. Because you're clearly designing your ethical system according to some *other* principle which *presides* over the cold, callous logic of aggregate-utility maximization. And if you're going to do that, you might as well do it openly.
P.S. the NAP is wishful thinking. The only real law is the Law of the Jungle.
> If you want to constrain Utilitarianism by the Non-Aggression Principle (NAP) and market forces, you're not *really* doing Utilitarianism, and you should probably make that clearer from the start.
Wow! I guess I really do have a clarity problem, because
1) I am NOT doing Utilitarianism, constrained or otherwise.
2) I have not advocated an NAP, and don't even know (or care) what it is.
My position is that Utilitarianism doesn't even make sense, except in the presence of a market and market prices, and that once a market exists, Utilitarianism doesn't even accomplish anything new, *except maybe to provide an evaluative framework for how well the market is functioning*.
It is a bit frustrating. I write a blog disagreeing with Utilitarianism, submit a comment promoting that blog, and get comments from one person saying that Utilitarianism is not to be disagreed with, and from another saying that it is not worth disagreeing with.
You are right. I really do have a clarity problem.
I mean, the phrases "collective utility maximization" and "AGGREGATE UTILITY GETS MAXIMIZED" pattern-match onto normative Utilitarianism, very strongly. VERY strongly. If you're going to say that and then *not* distance yourself from Bentham... ... ... sorry, but idk what you expected.
> [...] the end result, assuming free and voluntary exchange, is that an exchange rate gets established among the three currencies [...]
The NAP is a normative libertarian idea about being peaceful and non-coercive, so people can do (as you called it) "free and voluntary exchange". It's basically just "violence is wrong (except in self-defense)".
> My position is that Utilitarianism doesn't even make sense, except in the presence of a market and market prices, and that once a market exists, Utilitarianism doesn't even accomplish anything new, *except maybe to provide an evaluative framework for how well the market is functioning*.
lol, i have objections to this as well (though they're mostly the same objections I had earlier (i.e. market cap is a 2nd-order fiction)). But if I decide to continue down that route, I'll read the full blogpost and complain there, i guess.
Again, if you think that "market cap is a 2nd order fiction" is an objection to something *I* wrote, please read more carefully. Or explain how the objection works more clearly.
Maybe like, individual liberty is necessary to incentivize individuals to accept the social contract? If society can harm individuals in order to benefit the collective, those individuals are no longer incentivized to join the collective, and thus society breaks down. By protecting individual liberties, society ensures that nobody can be worse off than if they went at it alone, and thus society can function.
> Is there some justification for individual liberty which somehow overrules the utilitarian calculus of collective utility maximization?
I see two points in favor of individual liberty:
1) It feels good. (Yes, it's really that simple.) Therefore it should be included in the collective utility that is being maximized.
2) Instrumentally, in many situations individual liberty prevents us from horrible outcomes, when the leaders of the collective either do not have everyone's best interests in mind, or are too incompetent to achieve that efficiently (usually it's both).
If we get a benevolent superhuman AI overlord, it will still make sense to let people have as much individual freedom as possible without endangering everything (i.e. no freedom to play with nukes in your backyard), and maybe as a safety valve in case the AI turns out to be wrong about something.
The utilitarian calculus of collective utility maximization is an axiom. If you decide against it -- perhaps you conclude that the demandingness objection shows it to be too morally repugnant -- you are not overruling it but rejecting it.
"A better, more useful way"? If you argue that, you have already rejected utilitarianism because you are evaluating it as a tool to achieve some other system.
People are selfish. If you tell people they have to work in order to help others, they'll get lazy. If you tell them they can work to enrich themselves but they have to obey the rules of individual liberty and have to buy things from others instead of just steal things, everyone benefits. Once you have ASI, having people work isn't at all necessary, so it stops being an issue.
It would be nice if ASI maximized our welfare so we didn't have to bother doing it ourself. But I worry that won't happen automatically. So should you.
Does anyone work close enough to the consumer-facing parts of banking to help me understand why I'm currently two hours into opening a bank account for my [son's] cub scout pack, and it's still not opened?
On the one hand: small businesses are very different, scouting is legally weird in its own specific ways that are different from the way other small businesses might be legally weird, etc. So, I understand e.g. why I needed to come in person with my co-signer, and why the third human being who was meeting with us went to get approval from their supervisor.
On the other hand: two hours! With two of us sitting there!! And we're not actually done yet!!! How is that time actually being spent, and why does it require us to be at the branch in person?
This doesn't seem like a terribly isolated problem either; the reason I'm opening a new bank account now is that, for our previous one (with a different bank), we failed to accomplish the simple task of "transition me onto the account and the old treasurer off of the account" despite several long phone calls and an in-person visit. (In that case, most of the issue seemed to be that, despite having a *business* account, we kept being paired with *personal* banking agents.)
There are rather a lot of banks out there; why isn't the market solving this issue? I suppose most people just don't have to open that many accounts (let alone business accounts), and I get that marketing yourself as an "instant approval" place would not be ideal, but I'd like to understand what was actually happening behind the scenes while I was forced to twiddle my thumbs at the branch. "Willingness to sit there for two hours" doesn't seem like the signal that would demonstrates the scouting unit's trustworthiness??
Off the top of my head? Banks are pushing hard for everyone to online bank, so they don't have to pay humans to be working in branches and interacting with customers instead of trying to cut down labour numbers and get the remaining ones selling high-value crappy loans etc.
This also means they can close down low-performing (read: not in an area where they can sell enough high-interest crappy loans) branches and sell off the valuable real estate.
At least, that's what I take away from our own bank which now (in the name of efficiency and a better customer service) has caused it to be that instead of going to the local branch and depositing cash etc., now someone has to drive to the city where they graciously permit customers to transact in person. That's not convenient for the customers, but it's more cost efficient for the bank, so that's all that matters.
Which means that yeah, instead of being able to do things over the phone talking to a real human with agency, now it's all round the houses where nobody, if you can even get a real human, has the authority to do anything because now it's all "log in to your account and do this online". Except you *can't* do things like you describe online, so then all the account signatories etc. have to turn up and sit around waiting for the teller to ring headquarters and try and get through to business account manager to do this for you.
This generates the testable-in-principle prediction that it Used To Be Better -- that at some point in the past, opening a bank account didn't take several hours of personal time spread out over multiple days. I don't have personal experience with trying to open a small business account in the past, but I'm somewhat skeptical? Unless it used to cost more money and/or was less common for other reasons, so that in the past I simply wouldn't have had a bank account for this.
Presumably due to the Know Your Customer and Anti-Money Laundering provisions that were introduced by the Obama regime. I will assume that, if this is the case, someone like patio11 knows all the details.
(Banks have become odd these days: they're not doing lending and they don't want customers.)
I owned and operated a small business in the early 2010's, and it took me about 30 minutes to open a business account with the bank down the street. Of course, my business probably fit the standard template more closely than a cub scout troop.
My guess would be the supervisor is busy. They need the supervisors approval because "know your client" isnt automated for non personal accounts. Your contact has prepared the paperwork. The branches procedure mandate supervisor rubber stamps it. They might be the only person on site with a series 24 license. Which isnt strictly required but is the procedure theyve adopted.
So, in the spirit of figuring out "what would my life be like if I worked as the agent at the bank" -- what are *they* doing while my co-signer and I are discussing upcoming cub scout meetings instead of working? Tracking down paperwork for the next client? Writing a novel? Texting with their friends?
Were you opening up a not for profit account? Maybe that had something to do with it, like if they had to verify with the IRS and state revenue authorities that this was a legit non-profit and not some fake charity you and your friends were just going to use to launder money?
I'm not sure it will fully answer your specific question, but I found myself much less frustrated by banking inanities and much more grateful for banking after reading Patrick McKenzie's explanation: https://www.bitsaboutmoney.com/archive/seeing-like-a-bank/
Thank you! I work in software, and have occasionally needed to support old versions of the software, so I get that software is complicated, particularly if you need anything nonstandard. I'm genuinely more interested than frustrated by the specifics of "what was happening during the particular 2 hours that I spent at the bank" (with a side of "what part of that required me to be physically present at least at the beginning and at the end of the process"). I'm well-aware that the answer is going to be "something something software something something", but would like to fill in the something something's! I'm well aware that, whether or not they sound stupid, they would probably be tremendously difficult to change -- but that still doesn't answer the question of what is happening during that time!
Until recently when I transferred money on a Saturday from one bank to another it took a few days. I put it down to electrons being slow at the weekend, but turns out it was policy.
Insofar as ID checking is electronic, surely querying even several databases doesn't take hours? Insofar as it involves a human besides the one actually sitting their with us -- how do you actually involve a human in that?
In the US interstate-clearing used to take many days, sometimes two or three weeks. But I believe it is generally much faster these days, usually only a day or two. The exceptions are new accounts and large amounts, either of which may need further validation and hence delays.
In the UK check ("cheque") clearing used to be a two-step process, to and from clearing accounts in the Bank of England. This took up to five working days, and there was no incentive to speed it up because the B of E earned interest on the transient amounts, which they would otherwise have had to charge the other banks using these accounts. But these days, thank God, we can do free instant electronic clearing.
This is the first time I can remember someone being asked to not help anymore because they're *too* good at raising awareness due to their own popularity 😁
I posted this in the hidden thread and got no bites, will try one more time here in the open thread:
What follows is a startup concept I've been developing that may be of interest to people in this community that I am not in a position to build myself (due to lack of required skills, relationships, and available energy/time units). But I would really love to see it exist, because I think it has the potential to make a positive impact on public policy discourse and general Internet utility.
If you're interested in building this or know someone who would be, leave your email in a reply and I'll send you the detailed design doc.
EPISTEMIC ACCOUNTABILITY SYSTEM - BUILDING THE REPUTATION LAYER FOR PUBLIC DISCOURSE
I've designed a universal reputation system for public figures that measures two things: (1) Do you correct factual errors when identified? (2) Do you respond to serious challenges or ignore them? Scores are displayed via browser extension on any content anywhere online.
The mechanism: Anyone with public presence (bylines, social media following) automatically qualifies for paid jury duty on prediction markets, compensated per vote from the market vig. You lose jury eligibility if your reputation score falls below threshold - which happens when you ignore challenges or never correct errors. Prediction markets get expert signal and new users. We get validation (prediction accuracy tracked separately), revenue, and enforcement infrastructure.
This creates financial consequences for bad epistemic behavior while preserving intellectual freedom. You can be wrong frequently and maintain high reputation if you're honest about it. The system enforces: correct demonstrable errors, engage with serious criticism. If most public figures qualify for jury duty but keeping it requires epistemic virtue, ignoring criticism becomes professionally costly.
I believe this could transform public discourse in a very positive way, and also make someone a lot of money.
Let’s assume your technology even works in the first place; whose problem are you solving? The worst business models are technologies in search of a problem.
And when I ask whose problem you’re solving, get specific. Can you (even to yourself) name a person or a role who will get promoted for buying your product and fired for not buying it? If you don’t, even if you’re right about the technology you may not have a business.
Good question. I'm solving a customer acquisition challenge for prediction markets, which then allows me to make a small positive dent in the global public policy discourse problem.
Do prediction markets think they have a customer acquisition problem? Most of them are for wonks. A few (sports, stocks) are for normies. The ones for normies are popular (there are a lot of them) and the ones for wonks aren't as popular (there aren't so many of us). And if that analysis is correct, it at least suggests that the markets could be saturating the current norms and distribution channels; how are you really going to move the needle?
Sidebar: I've come to suspect that most people don't write for posterity, and would feel squeamish about signing up to be judged that way. I think most people either want to gamble anonymously for money, or with their close friends for clout. The ACX crowd is probably an outlier, writing with more of an ear towards self-consistency than the average person, but I don't know if that's enough to algorithmically validate appeals to authority. I'd also caution that this is catnip for Goodhart's Law; "I have the one true metric of prediction quality" is *begging* for an SEO-style industry of gaming the system, as Google experienced with PageRank. Finally, I'd point out that - as the old saying goes - past performance is no indication of future results. Cable missed streaming; not because streaming wasn't going to eventually become cable again, but because people hated the enshittification of cable. Who would have said they understood the cable/streaming market better than the cable guys? Right up until they didn't. And that's a pretty common story in technology; trains and cars, the internet, cell phones, social media, reusable rockets, lots of these technologies were poo-pooed by the old guard all the way to bankruptcy. How would your system account for *the underlying reality* changing?
To be sure, if you succeed I think that could be good for the world. I want more people to practice thinking like Galef lays out in The Scout Mindset. I just don't see how this model could get the traction to achieve that.
I'm confused. What is it that is being predicted in the prediction markets, and how are the topics/questions selected? How does this mechanism punish the ignoring of criticism and what makes criticism 'serious' enough to engage with?
Two thumbs up for good intentions, but yours is not the first attempt to systematize epistemic accountability. It is, IMO, a tougher problem than the one Satoshi 'solved' - with more moving parts to worry about.
"What follows is a startup concept I've been developing that may be of interest to people in this community that I am not in a position to build myself (due to lack of required skills, relationships, and available energy/time units"
If you don't want to see your own shit enough to overcome your limitations and make it no one else is going to be interested in making your shit.
You don't lack skills. Practice, almost certainly. However, you have access to the easiest to use, most capable machines that have ever existed in the history of mankind for virtually free. Don't have a skill? Think of the simplest, easiest, worst version of one thing that needs to exist for your project. Type into any of the text boxes at ChatGPT, Claude, or Gemini until that thing exists in some usable form. It will not be the form you intended when you started. That's unimportant.
Creating anything is completing a series of discrete tasks.
I do not have enough skill / patience with Premiere Pro to change the font size on text overlaid on a video. That wasn't a critical "skill" for making this: https://youtu.be/8opC-VYGiTc
Not amazing, certainly. Also not something that could have existed before AI, when my skill was the limiting factor. You have the choice now to be limited by the combined skill of You + A Large Number Of Highly Capable Models Specialized In Various Tasks. Lean on the stronger party in that collaboration.
All true. I'm just already building something that's requiring all my time and energy and ability to learn new skills. I'm also already a very heavy AI user. Thanks for the tough love anyway, it's good for the soul
Is there a word for the specific kind of assertive that cats get around mealtimes?
My foster cat doesn't meow in front of his food bowl or anything. He just sort of starts claiming my personal space as his. He sits on my computer mouse, or he moves between me and my keyboard, or he just sort of looms next to my arm and attacks my hand if it gets close to him.
And then it's my job to say "oh, I'll bet you're doing that because you're hungry," and give him some cat food, and then he goes away for a bit.
Sometimes I think I should get him some cat buttons that say "feed me" and "play with me" and "pet me". But if he's going to get adopted soon, maybe training him like that could lead to frustration later.
My cat does much the same sometimes, but her main tactic is to sit motionless by the food bowl and stare meaningfully at me or my partner, occasionally licking her lips!
I have noticed the main time she seems to be genuinely affectionate, without necessarily wanting anything besides stroking, is when she returns from outdoor patrols. I've concluded that this is largely instinctive in that each time she turns up after an absence she wants her presence to be noticed and acknowledged by the gang (us, her owners), in a sort of welcome ceremony, on the grounds that this is the friendly opposite of an undemonstrative reappearance which might be interpreted as stealthy and hostile (not that she probably consciously understands the reason).
On another topic, I've also long thought that cats like to vary their napping spots at regular intervals because they _don't_ like their own smell to be detectable, unlike dogs. So it causes cats stress if their spots are limited, unless coverings or blankets are regularly washed.
We have various terms for it; slinky, sexy-up, lovey. Typically displayed by a straight-up tail and rubbing their side against your calf, usually accompanied by purring. If I'm not in the kitchen, my current cat will sit near the bowl or fridge and stare at us until we see her. If I'm in bed, she will purr loudly near my face (not my wife's) until I wake. Failing that, she will snot me in the face, and as a last resort paw me in the mouth.
I feed her in the morning, and any other time she wants food. I don't think its worth the trouble to give her enough food at night to not wake me. For one, we try to feed her good canned stuff, expensive, and spoiling a bunch is dumb. We supplement with dry kibble in the middle of the night bc easier, but she loves it and would get fat plus bad kidneys. I don't feel bad about getting trained, I feel good about learning to communicate.
My cat is in-out, and a real brat about it. Her signal for out is a different one: she will stand by the door to the outside and meow until I let her out.
She has yet another signal for wanting to come in. There is a strip of plastic trim, a floor sweep, on the bottom of the storm door, and she will pluck it with yer claws as if she were knocking. It works great!
So we don't need buttons, our cat has come up with some of her own. I have never had a cat that communicated this well and specifically, unless I'm getting better at noticing as I age.
We call it being "bug reporty". The term started out with a very well-mannered cat of ours who would attract our attention, lead us to the situation he wanted addressed (usually a food dish he could see the bottom of, a water dish that was soiled or running low, or a toy he wanted us to use to play with him) and sit in front of it and look at us expectantly like he's just filed a bug report. Since then, the term has drifted a bit to encompass other cats who have less polite and less legible ways of asking for things.
There's definitely something to be said for giving them ways to ask for things they want. If you don't give them a polite way of doing so an reinforce it by being responsive to it, then they'll try other ways that you might not like as much.
A common failure mode is accidentally training your cat to ask for things in a way that is particularly inconvenience for you. The classic example is if you habitually feed your cat first thing in the morning, then your cat will start waking you up for food when you try to sleep in or even if they get hungry a little before you normally get up. If you respond to this by getting up and feeding your cat, you are training them to be a "alarm cat" who wakes you up at four in the morning.
So here's some view I sympathize with regarding NATO and the recent US security strategy that I suspect people disagree with but I'd like to hear the actual arguments.
The one-sentence summary is approximately "Europe's military underspending relative to the US and other allies was mostly not freeriding, it was mostly the US (and other allies) spending more than required for mutual defense, because they have more ambitious goals than mutual defense"
I'll give some reasons in the remainder of the post and then y'all get to tell me why I'm wrong :)
The premise is that NATO's main purpose is the defense of its members from approximately war.
I define a nation as freeriding on the alliance if they couldn't contribute their part to the required defense effort in a realistic attack scenario.
Since I believe this is central to my argument and people might have different defintions here's a toy example with some made up numbers to clarify what I mean by that.
If Europe was 30% of NATO (measured by GDP) then they'd be expected to contribute 30% of military power required in any realistic attack scenario. Let us say that NATO's combined military scores 1000 points and Europe only contributes 100 of these. On first glance this seems like freeriding.
And if NATO is attacked by some faction that also has 1000 military score and Europe only contributes to the defense the 100 points they have, then yes, I would contend that Europe is actually freeriding on the alliance.
However, if NATO is attacked by some faction that only has 100 military score and Europe contributes 30 points to defend the attack, then no, Europe is not freeriding on the alliance in this instance.
So the question really is, what are realistic attack scenarios and how much military power is required for mutual defense. Member states are free to have a military larger than this if they have more ambitious goals than just mutual defense. But that has no bearing on how much other members are required to contribute, the only thing that matters is that in the attacks that are actually realistic each member can contribute their part.
If that sounds reasonable then I would argue that for the most part Europe was not freeriding on the US, as they've mostly been able to contribute what was required for the attack scenarios that were actually realistic.
I think this is clearest in the post-soviet pre-crimea world, as there were no realistic would-be attackers at all. To argue freeriding you'd have to argue that there was a realistic attack scenario in which Europe wouldn't have been able to meet its commitments. If there are any that I'm missing feel free to tell me I'm wrong.
Post-Crimea you could argue Europe should've started the rearmament process, because you'd now have to assume that fighting with Russia was perhaps on the table again. But to argue freeriding you'd have to at least argue that Russia could've realistically attacked NATO at this point and that the European allies wouldn't have been able to contribute their fair share. See more on that last point below.
Post-Ukraine invasion I believe it's definitely fair to expect Europe to contribute their part against a possible Russian attack on NATO. But also we've seen how Russia actually performed during the Ukraine invasion.
It is my belief that we would've been ill-prepared had Russia magically sidestepped Ukraine and launched a conventional invasion against Poland directly, with massive ammonition shortages, long ramp up times on military production and many casualties.
But in my mind it's still clear that we'd win the conventional war, it would just be a lot more bloody and take a lot longer than necessary. And I think this is true still if the US or even all non-European NATO allies withheld any military support except what's their fair share (as measured by country GDP or some other metric).
And I don't know of any other realistic threat besides Russia that would require higher investment. Perhaps China could attack the US in the Pacific? I would say Europe would be ill-prepared to contribute there, but is that actually realistic?
In any case. It always seemed to me that Europe was not freeriding on NATO, rather they were spending little on military because there was no realistic threat to the alliance.
The Ukraine war demonstrates that everyone, even the US, has been underspending on their militaries.This is no great mystery, it’s called the Peace Dividend.
At the end of the Cold War, there was a euphoria that led to a global decrease in military spending. The Soviet Union had been defeated, and remarkably had chosen to accept that defeat with some grace. America and Europe cut their military spending, but Europe cut it by a lot more.
Now, nobody can afford to fight wars, even medium wars. The US and Europe couldn’t send Ukraine enough ammunition to drive off Russia. Israel and the US and friends expended a shocking amount of the their defensive missiles intercepting Iran’s missiles - and Hezbollah never even fully engaged before they were dismantled.
There are still a lot of flashpoints around the world, including some that are (in at least narrow ways) serious militarily contenders. Now, the US is short stuff like expensive missiles and aircraft carriers. Germany is short stuff like *machine gun ammo*.
In 1995 your argument was correct, Europe was spending appropriately to the threat Russia posed. But they didn’t wake up Russia interfered in Chechnya. They didn’t wake up in 2014 when Russia invaded Crimea. They have now, since 2022, woken up. They would have saved a lot of money if they’d woken up sooner.
Thanks for the link to the Perun video, I've watched it and it's part of what got me thinking that maybe the argument I'm defending here isn't obviously wrong.
So I'm not sure what we actually disagree on.
It seems you agree that it was rational to cut spending after the Soviet Union collapsed.
I agree that the ideal time to ramp up military production was maybe 10 years ago and that it's cost Europe a lot of money to do it now all in a hurry.
I guess my contention is that even in this world where we failed to ramp up as early as optimal, Europe and non-US NATO did not disarm to the extend Russia could've taken them on in a conventional war.
At least my takeaway from the video you linked is, that if Russia invaded Europe and the US would've stayed out of it entirely, then Russia would probably still lose, though it would be extremely costly and bloody for NATO.
Which is a stronger claim then what I'm defending here, which is that NATO could defend Russia if the US committed exactly their share to NATO and not a cent more.
I think you yourself highlighted what we disagree on, right at the end. In my opinion, the correct level of NATO spending isn't to win a war, but to deter a war in the first place.
Because actually fighting a war is much more expensive than deterring a war. You have to buy a lot of stuff in a hurry, you use up your equipment so you can't keep operating it or sell it on the second-hand market, and (most importantly but least visibly in the short term) a lot of young people die. To use a historical example; the Axis in WWII underestimated US production capability by an order of magnitude, and was therefore willing to fight the US. If the US had been visibly more powerful, Japan might not have gambled on Pearl Harbor and instead left China in the early 1940s.
In the deterrence respect, Ukraine has been a very mixed bag for NATO. It has shown Russia and China that war can be very costly ... but that the West isn't necessarily better at this kind of war. Most of NATO's advantages (airpower, combined arms maneuver, intelligence, electronic warfare) aren't on display. Russia's "bring them down into the mud and win at mud wrestling" theory of war seems at a surface level to have some merit, even though just as clearly they can't win an outright modern conventional war.
But NATO isn't innovating faster than e.g. Ukraine in the drone domain. NATO isn't producing enough shells to feed the guns. NATO has been slow to establish the kind of long-term, large-buy ammo contracts that would get them out ahead of shell hunger for next time. Europe did solve their own natural gas shortages by (in Perun's words) turning on the money spigot. All taken together, if you're China and you're thinking about taking over Taiwan, your takeaway might be that Europe and America are weak, selfish, and short-sighted; full of unrealized potential but not able to stop a cohesive enemy who strikes decisively. That is not a success of deterrence! Sure, it's likely over-learning the lessons of Ukraine, but organizations over-learn all the time.
The current funding level of NATO has, in my opinion, made war more likely in the short term. Even if it's a war NATO would win, that's still a failure of NATO at its core function.
> I think you yourself highlighted what we disagree on, right at the end. In my opinion, the correct level of NATO spending isn't to win a war, but to deter a war in the first place.
I think this is basically reasonable, but also much more impossible to answer. So I chose to discuss the question that lends itself better to discussion.
For one, NATO has so far successfully served it's purpose and deterred Russia from attacking NATO.
But also it is unclear to what part the military spending of particular nations deters Russia or any other potential adversary.
I think it's safe to assume, that Russia is a lot more scared of the US Army than of the Bundeswehr. But it's pretty unclear how much Russia would be willing to attack NATO ex US. I mean, France and GB have nukes as well. And even in a conventional war, we just do not know if Russia would consider going at it against what non-US NATO is able to field. And if they are unwilling, we're likely never going to know the reasons and how deterred they felt by European militaries.
As for what happens in the pacific, I'm not sure this is that much about NATO or Europe at all. Other commenters have made the point that whatever happens there is not covered by NATO treaties, so Europe would not be bound to help. In that respect it's unclear how European military spending has any impact on Chinas decisionmaking process regarding Taiwan.
I should say I did not double check if NATO not covering the pacific is actually true.
NATO mostly doesn't cover the Pacific ... but they are definitely more friendly to and likely to help out some countries there than others.
What I believe would deter China - and Russia - is showing that Western nations can effectively and cost-effectively win modern warfare. And we ... have not really out-innovated or outproduced Russia (or Ukraine) in the things Ukraine needs. The thing that Ukraine has been screaming that they need, from the beginning, is ammunition. They need shells and drones and missiles. And we, in theory, could be tooling up and producing them. Arsenal of Democracy style. But it's taking shockingly long. We've had some successes, but honestly a major share of the munitions that are holding the line are made in Ukraine with Chinese components.
And if you're China, who can and does produce things at enormous scale, that has to be a huge encouragement to see that nobody else can even sort of match you when motivated.
The general thrust of the argument here is that they dont have to be as martial as the US to achieve their obligation. This part may be true, but your reduction to points smooths over some bits
To achieve the mutual defense principle NATO members have to be able to do two things
A) hold there own long enough for there to be something left for the others to defend
B) project enough power to come to the defense of others
The plausible max threat is invasion from Russia. That's how it started thats how its going.
This makes A much harder for some than others. Baltics in particular would need more of their GDP to put up a self defense. They generally do. It remains to be seen if they could.
B requires individual countries to be able to power project. Its not okay for them to only be able to defend themselves. Thats where you treating the European countries as a single block misses the problem.
Ukraine is a test run in their ability to defend another country. If France cant come to the defense of Ukraine, it cant come to the defense of Poland. On the whole Europe is failing this test. They have contributed more than the US proportionately but this comes in the form of buying US weapons and giving them to Ukraine. Their own militaries seem to have very little to directly offer. They have not been able to achieve their own stated objectives so Im not inclined to believe if it *was* Poland they would do much better.
The biggest reason that France can't ensure the defence of Ukraine is that they can't / don't want to send their military, since Ukraine isn't in NATO. Provision of weapons and other support is completely irrrelevant in relevant importance.
> If France cant come to the defense of Ukraine, it cant come to the defense of Poland.
Well, yes, they can, in the form of, you know, actually _coming_ to defend them.
Whether European NATO members will actually honor their obligations, especially if rules by right-wing-populist governments, is another question. But the parallel between Poland and Ukraine that you're asserting isn't there.
Not sure I'd describe Ukraine as a fair test of NATO. Since Ukraine is not a NATO member, NATO can't directly defend Ukraine. If NATO could send forces directly to Ukraine instead of just selling weapons, I think that would change the equation a bit.
In a world of uncertainties it is a data point. And I think the collective evidence, including Afghanistan, is not great for european countries, non UK in particular to be able to project power which they have to do to defend a member state. I have a lot of uncertainty on this point. I would be happy to see a strong argument that they could. But "theyd do better if they had to" isnt so reassuring
NATO certainly *could* defend Ukraine directly, just as it defended the citizens of Kosovo, Bosnia, and more recently Libya, in spite of those nations not being members of NATO. The victims du jour not being members of NATO, means that NATO has a choice. They have chosen. It remains to be seen whether they chose wisely.
"It is my belief that we would've been ill-prepared had Russia magically sidestepped Ukraine and launched a conventional invasion against Poland directly, with massive ammonition shortages, long ramp up times on military production and many casualties"
Well thats the whole problem isnt it? Poland borders Belarus and the Baltics border Russia, no magic required. The Ukraine experience demonstrates European states were under prepared for a realistic invasion scenario.
I think I'm right in saying that until recently the US was fairly relaxed about European countries shirking their full obligations to NATO, on the grounds that the EU might otherwise one day become a threat itself to the US, so their relative military weakness was actually reassuring. Twenty or thirty years ago some futurists were confidently predicting that the US and EU would be at war by now!
Imagine a parallel universe where European countries are spending on military as much as USA.
Then Trump makes a joke about invading Denmark. The Danish respond by deploying their troops next to the border, just in case. Trump sees this as escalation and deploys more troops. ... A few weeks later, the nukes start flying.
(Or maybe in that universe Trump wouldn't be making this kind of jokes, but I wouldn't bet on that.)
In such a universe, Trump is making different jokes, because "Denmark isn't carrying its weight militarily" isn't a running gag someone can make jokes from.
Also, Trump is probably making those jokes at another one of his dinner parties or perhaps backstage at another episode of The Apprentice, because Americans had a very different set of concerns that didn't suggest "let's see how this Trump guy does if we elect him".
The United States is a loose cannon engaged in constant aggression, and its pushing up "defense" spending targets has more to do with revenue streams for politically connected, US based arms manufacturers than any actual defense.
Who are Europeans being "protected" from? Russia, which has struggled mightily to take a portion of Ukraine?
> Russia, which has struggled mightily to take a portion of Ukraine?
That's hindsight. After the Russian blitzkrieg failed, it turned out that Russia was not sufficiently prepared for a long war. But if a few things had turned out differently, Kyiv would have been conquered in the three days, and then the situation in Ukraine would be dramatically different.
If you roll a die and you get 6, you have the benefit of insight, but you should not base your policy on the assumption that you will always get 6.
My model is that the blitzkrieg on Ukraine was a coinflip. We could discuss specific probabilities, but to keep it simple, let's just say they were somewhere between 20% and 80%.
Option A: Russia successfully takes over the capital, and afterwards spends a few years cleaning up the remaining centers of resistance. The West will send a few angry letters, but... it's already done, you can't kick Russians out of Ukraine without sending your troops which is not going to happen, and the pockets of resistance cannot survive against the meat waves mostly consisting of forcibly recruited Ukrainians from places that Russia already has firmly under control. Russians will pinky swear that this was a unique situation and they would never do anything like this again (and for the next 5 years they may even keep their promise) and the Western politicians are happy to accept the excuse to lift all sanctions.
Option B: Russia fails to take over the capital, and gets into a prolonged war where it controls a minority of the territory, but faces the organized Ukrainian army supported by occasional weapon deliveries from the West. The most likely outcome is Russia taking a small part of territory.
Russia got the worse option. Yet, how bad is it, actually? If at some moment Ukraine gives up a part of its territory and the war stops, then... Russia has a new piece of territory, the West will be happy to remove the sanctions soon. Yeah, lots of tanks were destroyed, but as long as Russia can keep selling oil, it will rebuild them, plus it will build a new army of war-tested drones. The loss of soldiers will be replaced by kidnapped Ukrainian children, brainwashed and adopted to Russian families. Ten years later, Russia is as good as new, and can try the same thing in another country, because why not, if the only options are "win" and "win less".
I am not saying that Russia could take over the entire EU, but it could try the same kind of blitzkrieg e.g. in Estonia, in a few years. Only now the skies will be full of drones since the first day of attack. What is EU going to do about it, besides sending angry letters? Are German or French soldiers really willing to die in Estonia? (Heck, will EU even be able to send the angry letter, if Hungary or Slovakia or anyone else will veto it?)
So, on one hand, Russia is unable to to take over the entire EU. On the other hand, if it goes like "hey, we just take this small piece and stop there, pinky swear", EU doesn't really have a good answer to that. Especially if the territory is taken over in three days, which seems technically possible. (It failed in Kyiv, but only because Russia had logistical problem getting the tanks to the capital. This time it will be drones instead.)
First, I'd like to make a case to stay with the imperfect measurement of %GDP spent on military. It is what every President has used to scold NATO partners, so that's a place Trump has been somewhat conventional. It is a flawed metric, tho, bc if your economy is strong and your GDP grows, you can fall behind on your military spending without materially changing your readiness (and totally rewards countries with poor GDP).
Still, it is the one we have, and I believe as of July 20+ countries are above the 2% target, with 9, including Canada, Spain and Italy lagging.
This is IMO a strong argument against overall freeloading, but a clear example of some countries coming up short.
I credit Russia's invasion of Ukraine with helping change the military investments. Trump did a lot of hectoring, but if he gets any credit at all, it is only in the backhanded way of allowing/encouraging Russia's aggression, and his tepid--and at times extractive--support for Ukraine.
As for the rationale of NATO, it has always been mainly about USSR/Russia. To this end, Ukraine has done an exceptional job of holding back Russia's military. Verily, I commend them to NATO as the ideal future partner! They fight like hell, and they invest greatly in the defense of Europe's NATO partners, albeit for selfish reasons
Those who painfully follow the stated policy of Trump 1&2 re NATO will see that the goalposts have inexorably been moved farther back over time. First, you have to pay your way. Then it was 'we might not defend you if you don't.' Now it has fully morphed into 'we don't think we need to defend you in any event.'
Finally, for the record, there has only been 1 article 5 activation, and that was 9-11. NATO partners fought and died beside us in Afghanistan.
You may recall that many NATO countries felt that Iraq did not meet the article 5 threshold. This caused much resentment in the US (remember Freedom Fries?) but has turned out be the correct call. Not sure why I felt the need to give a history lecture, perhaps it was the hypothetical that the US could be attacked and European partners not come to our aid. It happened, and they did.
Re China attacking the US in the pacific, this is possible in the leadup to e.g. Taiwan Strait war, in which case it would be expected to happen between 2027 and 2030, but NATO doesn't cover US assets west of North America, so joining in would be a choice not a treaty obligation.
It certainly covers US assets west of California: it covers all of North America, which includes Alaska, which is pretty far west of California.
Now China would have to be pretty dumb to strike US assets in Alaska or the rest of North America, but if things escalate to the point of outright war with the US then it very well could happen.
The Delian League, also known as the Athenian Empire, was a collection of Greek city-states largely based around the Aegean Sea which operated under the hegemony of Athens. This alliance initially served the purpose of coordinating a united Greek front against a perceived looming Persian threat against the Ionian city-states which bordered it.[38] The members of the Delian League were made to swear an oath of loyalty to the league and contributed mostly monetarily but in some instances donated ships or other forces.
Benefitting greatly from the influx of cash coming out of the 150-330 members,[39] Athens used the money to reinforce its own naval supremacy and used the remaining funds to embellish the city with art and architecture.[38] In order to maintain the new synoecism, Athens began using its greatly expanded military to enforce membership in the League. City-states who wished to leave the alliance were punished by Athens with force such as Mytilene and Melos. No longer considered her allies, Athens eventually began to refer to the members of the Delian League as "all the cities Athens rules."[38] Athens also extended its authority over members of the League through judicial decisions. Synoecism under the Athenian Empire was enforced by resolving matters of and between states in Athens by courts composed of Athenian citizens and enforcing those decisions through the Athenian military.
See, now that's how you run a military alliance. The guys doin' the heavy lifting and taking all the risk get to set the terms. I say NATO should acknowledge the facts on the ground and rename itself the Brusselian League (NATO headquarters are in Brussels), and the US should use it to forcibly extract resources from member states.
You forgot to finish the story! The best part! The part where the constant confrontational situations the Delian league required of Athens led to the inciting incidents of the ruinous Peloponnesian war where several of their “subjects” revolted. It was the beginning of their decline. And when they reformed the Delian league a second time, they were very careful to learn their lessons from the first go around and treat their partners better. Thucydides can be dry at times, but skip him at your own peril if you want to opine:
"The growth of the power of Athens, and the alarm which this inspired in Lacedaemon, made war inevitable"
There is no modern equivalent of the Peloponnesian League, so we can safely disregard this particular lesson of history and exploit Europeans to our heart's content.
I'm joking of course; this was a mere modest proposal. But is it correct to say the war was a product of confrontation *required* by the Delian League? I was under the impression it was more due to Athenian mistakes in foreign policy; misreading the goat entrails, so to speak, and overestimating their own military prowess relative to Sparta.
You are of course right in the broad strokes but the war and subsequent spoliation of Athens doesn’t seem inevitable in a counterfactual non-Delian world. It may be that I take Thucydides too naively at his word, but the Potidean affair is an explicit inciting incident where Athens’ tendency to treat her de jure allies as de facto lesser subjects led to hostilities
If we're worrying about word forms, "Delian" already appears to be malformed. Wikipedia notes that the organization was referred to at the time as "the Athenians" (or rather, I assume, as the Athenaioi - "Athenian" is just as bad as "Delian"); Greek wikipedia doesn't use an adjective at all, calling it "the alliance of Delos", but it does provide an adjective for the Peloponnesian War, and as you might expect it's "Peloponnesiac".
If you're curious, think of amnesia / amnesiac. Mania / maniac is the same thing, except that we made "maniac" a noun and then needed a new adjective; "maniacal" uses a Latin suffix.
The use of -n- to derive an adjective meaning "related in some unspecified way to the root word" is specific to Latin. Greek is likely to use -k- (compare -ic, though that appears to be a morpheme that Greek and Latin share).
English theoretically uses the same -k-, although in our case we see the progression of sound changes -ik- → -ig- → -iy- → -y, leaving us with e.g. the suffix of "messy".
The problem with "Athenian" is that you don't want an adjective meaning "related to Athens in an unspecified way". Athenians are related to Athens by the fact that they come from Athens, and for that relationship in Latin you use the gentilic suffix -ensis.
Here's the thing: NATO is not really based on mutual defence at all. From the US's point of view it's not a sensible, rational alliance whereby the US defends Europe against Russian invasion and Europe promises to defend the US against, like, Mexico or something, because the latter scenario is implausible.
The US defends Europe, by which we mean the existence of Europe as a collection of civilised democratic countries, because it wants to. It has repeatedly done so, at great cost, even when it's not obviously in its own interest.
It's less of a relationship between two equal partners and more like the relationship between two brothers, one reasonably successful and one an unemployed bum. The employed brother lets the unemployed one live in his spare room in exchange for doing some chores around the house, not because this is an equal exchange but because he wants his brother to get his life together. And the unemployed brother keeps shirking his chores and refuses to look for a job, and is now making up arguments as to why he's not _really_ a free rider.
It seems to me you're redefining NATO because you believe NATO as defined by its actual treaties is not in the rational self-interest of the US.
And then you're arguing that Europe is freeriding on what you believe NATO should be.
But I'd much rather argue the actual treaties than our personal interpretation what the treaties should or shouldn't have been.
If the US shares your sentiment and believes there's not enough value for it in NATO, then they are free to leave the alliance, they can actually do that.
Ah horseshit. You skipped the part where Europe was the US' most important trade partner, and where Europeans spent a shitload of money on defense while the Soviet Union was an actual threat, and the part where European partners supported the US in Afghanistan, and so on and so forth.
Yes, Europe underestimated the threat that Russia again started to pose. So did the US. But, reducing military spending after the Soviet Union had ceased to exist was not "free riding", it was the only rational thing to do (up to the point where it wasn't anymore, and we missed it).
Case in point: The US also reduced military spending (in terms of GDP) dramatically after 1991, and only ramped it up again after 2001. But not because of the threat from Russia (or China), but to fund their excursions into the Middle East.
It's a sensible alliance because the goal was and is to avoid another brutal, costly war in the heart of Europe that might suck in the US again. It was as much about that as about defending against outside threats.
I think that threat is gone (nobody thinks the Germans and French are going to come to blows), but the US also really doesn't want any other European powers to get nuclear weapons - and a situation where the US pulls back from the alliance in a serious way is one where you probably have another half-dozen nuclear weapons-possessing countries in 5-10 years.
That's a truism only in that everything that happens in the world somewhere has impacts, positive or negative, on America.
Europe was the birthplace of The Enlightenment. With all of the attacks on speech, property, religion, and other American values, they are drifting further and further away from those values. We shouldn't be spending blood and treasure on them.
The big problem is that if the US kicks the Europeans to the curb, American High Society people won't get invited to the swanky parties any more. And it's worth all the plebeian tax dollars and lives to defend that access.
With all the attacks on speech, free trade, freethinking, racial equality democracy and other American values , America is drifting further and further from American values.
How do you go from "massive ammunition shortage and long ramp-up times on military prodiction" to "we'd win the conventional war"? Is the plan to call for a time-out when you run out of ammunition, with Russia patiently waiting for you to build more?
If the plan is for the United States to provide 970 points of extra ammunition to beef up your 30-point military, then yeah, that's freeloading.
To address your point, my assumption is that there's ammo stockpiles in every NATO country and that in a hot war event we would ramp up ammo production and consume from our stockpiles in the meantime.
Presumably ammo would need to be rationed, perhaps right from the get go, to buy more time for production ramp up.
But it seems to me that NATO, even NATO ex US, has the industrial capacity to outproduce Russia, provided it can last long enough to make that count.
So in my mind the the loss scenario is mostly that Europe falls to Russia before it can make use of its economic advantages. Do you disagree with any of that?
Free riding means benefiting from a shared resource, public good, or collective effort while paying for or contributing to its cost less than is fair or required. By this definition Europe was free riding because it benefitted from security and failed to meet its treaty obligations. In fact by this definition Europe is still free riding.
You're describing why it's rational to free ride. Basically, the US would keep global security and Europe could not help with that and only help with its immediate threats. See, for example, how France has repeatedly tried to undercut the US with China. Even after the Ukraine invasion and under Biden France insisted on saying it would not help the US defend Asian democracies.
> By this definition Europe was free riding because it benefitted from security and failed to meet its treaty obligations.
It's open to debate. Western Europe failed to meet obligations and retained membership in NATO, which can be characterized as a benefit. But the actual value of that membership can be questioned.
Eastern European members behaved differently. They retained membership in NATO and they also met the membership requirements in full. I don't think it's a stretch to say that that was because they saw more value in membership than Spain did.
It's Poland, the USA, Greece, Finland, Estonia, Latvia, Lithuania, Slovakia, Hungary, Romania, and England. Or to put it another way, it's every country that borders Russia, Belarus, or Ukraine, plus the USA, England, and Greece.)
If you take Europe as a whole they failed however. But I take the point that specific nations did not, though even most of them only after 2014. But still, I give them credit for that and would prefer the US was higher resolution.
Europe is free riding global security as much as China, Japan, Brazil, Saudi Arabia or anyone else does. US is not doing this for Europe but to further its own interest as global hegemon. Until the 20th century the US was free riding the global security maintained by the British Empire the same way.
On the other hand, as its name literally defines, NATO is treaty for the North Atlantic region. After 9/11 European allies helped the US in Afghanistan, the attack having been in the specified area. However, during the Falkland war US did not intervene directly as that not included in the area.
If there is a war against Russia in Europe, would you expect Taiwan, South Korea and Japan to contribute forces? If they are not inclined to do that, are they also free riding their alliance with the US? How about the gulf states that host US military bases, were defended by the US in 1991 and most likely would receive US help in case of a war with Iran? Are they expected to help against China and Russia if it comes to it?
> Europe is free riding global security as much as China, Japan, Brazil, Saudi Arabia or anyone else does.
We don't ship weapons to China or aided them in securing their borders or were giving them military assistance in Myanmar. Nor equivalently for Brazil. We give Saudi Arabia a lesser amount of help because it's not an ally. Japan you're right about. Though it's still slightly lesser because, unlike Europe, they didn't agree to certain contributions then ignore them.
> After 9/11 European allies helped the US in Afghanistan, the attack having been in the specified area.
Yes. And the US has given more to Ukraine than the EU gave to the entire war on terror despite the US having no treaty obligation to Ukraine. Therefore, why do Europeans have any objection to the US pulling back? Because alliances are more than legalism.
> However, during the Falkland war US did not intervene directly as that not included in the area.
The US did, however, back Britain and supply refueling, intelligence, weapons, and Britain refused further help.
> Until the 20th century the US was free riding the global security maintained by the British Empire the same way.
The British Empire was a security threat to the United States, not a net contributor, until after WW2.
> If there is a war against Russia in Europe, would you expect Taiwan, South Korea and Japan to contribute forces?
Yes.
> If they are not inclined to do that, are they also free riding their alliance with the US?
Yes.
> How about the gulf states that host US military bases, were defended by the US in 1991 and most likely would receive US help in case of a war with Iran? Are they expected to help against China and Russia if it comes to it?
"The British Empire was a security threat to the United States, not a net contributor, until after WW2."
China would also describe the US as a security threat for the past 80 years, yet the rise of China would have been impossible without global trade security maintained by the US.
In fact, the US supplied Chinese forces during their war against Japan and sanction Japan massively due to its war on China that time. (a lot like the Ukrainian-Russian situation). Without that, the outcome of the war could have been very different.
"And the US has given more to Ukraine than the EU gave to the entire war on terror despite the US having no treaty obligation to Ukraine."
The US did not help Ukraine to help Europe. US helped because
1. The Russian invasion/intervention was against the rules based international order that was primarily created by the US.
2. The invasion was aimed specifically to reduce/constrain US influence in Ukraine.
3. Russia was already actively pushing against US interests in multiple areas and a victory against Ukraine would have made it a larger threat even in those other areas.
"We don't ship weapons to China or aided them in securing their borders or were giving them military assistance in Myanmar. Nor equivalently for Brazil. We give Saudi Arabia a lesser amount of help because it's not an ally. "
In 1991 the US deployed forces to defend Saudi Arabia from a potential invasion, and helped freeing a Saudi town Iraq captured with actual American boots on the ground, and was the primary force that smashed the Iraqi army and liberated Kuwait. This is certainly larger contribution than what Europe got since WW2.
"The US did, however, back Britain and supply refueling, intelligence, weapons, and Britain refused further help.
....
Therefore, why do Europeans have any objection to the US pulling back? Because alliances are more than legalism."
I am pretty sure Europe would provide such too. In fact, most likely some countries (primarily the UK, and to some extent Poland) would send fighting forces. But that is just the same as US contribution in Ukraine (or in Kuwait, Vietnam, Syria, Israel): the US did that due to its own interests (and to some extent friendliness) without having any legalistic demand. Likewise, NATO in the context of a pacific war has no obligations to help, except for interests or being friendly. (note, though some NATO countries did help largely in 1991 and 2003, and most of them contributed to the security forces after the Iraq invasion; lots of them also helped in the fight against ISIS and do not provide weapons the US embargoed states).
I agree, the US has technically no obligations to help Ukraine. The EU has no obligation to do that either. And while the Russian framing was primarily about US being the big bad guy (at least until Trump got reelected), I think the push on Ukraine was more against the EU and it is certainly more of a direct threat there. I do agree that Europe should have been able to arm Ukraine alone and in any ways should have done larger part of the lifting. That being said: The previous US administration did make promises to Ukraine about supporting them which influenced Ukrainian decisions about conducting the war. Going back on this for reasons entirely outside the influence of Ukraine (such as 3rd countries not spending enough on defense) would be a dick move (and as hinted before, Russia winning would be actually counter to US interests.).
I think Taiwan, Japan or the Gulf states have a very different understanding about sending forces to a potential war against Russia (or in the case of the latter, against China too), and their median citizen or even politician would be surprised and less likely to do that than most Europeans.
> China would also describe the US as a security threat for the past 80 years, yet the rise of China would have been impossible without global trade security maintained by the US.
Yes. And if the British had had primary responsibility for policing American trade, and had not started wars or attacked the United States, the situations might be symmetrical. Likewise if Britain had helped the US rather than selling weapons to the Confederacy or other nations the US was at war with or providing them diplomatic cover. Britain was much less benign to the US in the 19th century than the US was to China.
> The US did not help Ukraine to help Europe.
You are now arguing intentions to get around the simple fact that the statement is true. The US has done more to help Ukraine than Europe did to help in the war on terror despite Europe being treaty bound and the US not being treaty bound to help Ukraine. I could equally argue Europe got into the war on terror not to help the US but to prevent, and later avenge, their own terrorist attacks. Which did happen. I don't make such an argument though because I'm consistent about this.
> In 1991 the US deployed forces to defend Saudi Arabia from a potential invasion, and helped freeing a Saudi town Iraq captured with actual American boots on the ground, and was the primary force that smashed the Iraqi army and liberated Kuwait. This is certainly larger contribution than what Europe got since WW2.
Saudi Arabia allowed US troops into the country to help with the planned liberation of Kuwait which was an ally. And the idea Desert Storm was a bigger contribution than what Europe got is absurd. First, it was just monetarily less costly (though with inflation more costly) than Ukraine has gotten. Secondly US forces in Europe have been there longer over the entire period such that it adds up to a much bigger commitment. There are US troops in Europe right now to stave off invasion.
> I am pretty sure Europe would provide such too.
While I disagree with a fair amount of specifics, I agree Europe feels the US has been behaving badly. And in some ways are correct. However, Europe is behaving worse than the US even under Trump. Though the US is by no means perfect or innocent.
> I think Taiwan, Japan or the Gulf states have a very different understanding about sending forces to a potential war against Russia (or in the case of the latter, against China too), and their median citizen or even politician would be surprised and less likely to do that than most Europeans.
I disagree. The point about Russia is particularly ill taken considering Japan and South Korea have disputes with Russia and have been threatened by them. A fact the Europeans neither seem aware of or interested in.
"Britain was much less benign to the US in the 19th century than the US was to China."
I dispute this, although such things are difficult to measure. However, let's not ignore that the US supported the losing side during the Chinese civil war, and in fact, still supports and basically guarantees the independence of what remains of it. And of course there was the Korean war with hundreds of thousands of Chinese casualties.
"You are now arguing intentions to get around the simple fact that the statement is true."
Okay, so the US supported a 3rd country more on its own accord than Europe supported the US via a treaty. What do you want to imply with this? Europe is bad because the US was generous with someone else?
"Saudi Arabia allowed US troops into the country to help with the planned liberation of Kuwait which was an ally."
The first group of troops arrived to Saudi Arabia to help against potential Iraqi invasion.
Kuwait was not an US ally at that time, just a generic friendly nation. The US had no legal responsibility to help them at all. Even less than Ukraine if I may argue, as there had been no US president who tried to add it to a US lead military alliance and Iraq was not supporting the enemies of US at that time.
" And the idea Desert Storm was a bigger contribution than what Europe got is absurd. First, it was just monetarily less costly (though with inflation more costly) than Ukraine has gotten. Secondly US forces in Europe have been there longer over the entire period such that it adds up to a much bigger commitment."
They have been in Europe longer, but there have never been as many US troops in Europe (since WW2) as what were deployed for the Gulf War, and they certainly did not have to march in and die while liberating a NATO country.
Through the decades the monetary cost likely have been more, but US troops did not have to actually fight and shed their blood. This is not an apples-to-apples comparison though, so could be interpreted either way.
Anyway, the main point I am arguing for is not that US did not spend a lot of resources on European defense, certainly more than what Europe contributed in the other direction.
My thesis is that the US had been doing that all around the world through much of the 20th and 21th century, and Europe is not an outlier in receiving that. The US was the primary force in Korea, then later in Vietnam and of course through multiple wars in the Middle-East (and with smaller scale but clear consistency even in other parts of the world). US still deploys lots of forces all around, including Japan and Korea and a bunch of Gulf countries, and would likely intervene on the side of many others if needed (Taiwan, Israel), and is doting out a lot of free military support (e.g, Egypt receives over a billion USD per year).
If anything, Europe helped more in US adventures than other countries. E.g., British, French, Italian contribution in the Gulf War, bunch of countries contributing during the Korean war, British and Polish during 2003 Iraqi invasion, a whole bunch of European countries contributed to bombing ISIS in Iraq/Syria, the UK and France doing their own policing missions in Africa, and even now, it is basically only European Navies that are trying to help the US in keeping open the Red Sea.
On the other hand:
-Kuwait and Saudi Arabia did not provide any help in Afghanistan, they did not even participate in bombing ISIS or during the Red Sea policing, which was/is next door and direct threat to them. They are also more in line with China than Europe.
-Japan did not help in Afghanistan either and if they were opportunistic against Russia, they would likely contribute weapons to Ukraine, which they do not do.
-Israel and Taiwan did not provide any direct military help to any US endeavours to date as far as I know.
Could European countries be a better allies? Sure. But a lot of them are already among the best the US has, and certainly the worst ones are not in Europe.
"However, Europe is behaving worse than the US even under Trump."
Which particular behaviour was very bad from Europe?
Whatever it is though, I doubt it would be worse then trying to bully other countries to become part of the US or cede territory to it via economic and diplomatic pressure and even hinting the possibility of military invasion in case of refusal. Even worse, these were not against free riding countries, but against Canada and Denmark, both with very strong integration to US defense and consistently punching above their weight in contribution during military matters.
Is it your position that Europe was free riding of NATO or the US established global world order?
It is true that Europe was not meeting the minimum spending mandated by NATO treaty. It is my position that 1) they should've followed the rules that existed (i.e. spend the required minimum on defense) and 2) they should've changed the rules (there was no rational reason for NATO to spend 3% GDP for mutual defense).
If the US wants others to contribute to global security they should establish a coalition for keeping global security, to the best of my knowledge NATO does not have that purpose and it is unreasonable to expect member states to pay for B because they are in alliance for A.
But I feel the argument for freeriding on NATO must be based on Europe not contributing enough for mutual defense, not based on Europe not contributing enough for global security.
Edit: But also in general I find the argument "I've decided to do something that's good for both of us, so now you better pay up" to not be very convincing.
“It is true that Europe was not meeting the minimum spending mandated by NATO treaty.”
Actually, the NATO treaty doesn’t mandate spending. At each NATO summit the members agree on a declaration, but those declarations aren’t treaties. Here is the most relevant portion of the 2014 NATO summit declaration:
“Allies whose current proportion of GDP spent on defense is below this [2%] level will: halt any decline in defense expenditure; aim to increase defense expenditure in real terms as GDP grows; [and] aim to move towards the 2% guideline within a decade with a view to meeting their NATO Capability Targets and filling NATO's capability shortfalls.”
So not a treaty and not a promise to actually get to 2% of GDP. Not all NATO members met the 2% guideline “within a decade,” though they are expected to this year.
The 2025 summit set a 3.5% target (as well as a 5% target using a broader definition of defense) to be achieved by 2035.
I'm not sure what distinction you're trying to draw. There's a public good here and whether you define it as "global security" or "security in Europe and North America" doesn't seem especially salient to me. In either case, Europe was not contributing enough for either.
My position is: if you make an agreement, the other side keeps it, you do not, and you benefit from them keeping the agreement then you are free riding. If Europe didn't wish to make the agreement it was free not to and to bear the consequences of that.
I am easier on Japan because they did not make such an agreement. So while I think they also acted in ways that weren't ideal they get off easier than Europe. I'm also especially frustrated with Europe's constant attempts to directly undermine American security and hypocrisy on the economic and political front in ways that don't really apply to Japan (but do to some extent apply to South Korea).
> My position is: if you make an agreement, the other side keeps it, you do not, and you benefit from them keeping the agreement then you are free riding. If Europe didn't wish to make the agreement it was free not to and to bear the consequences of that.
Thank you, this helps me understand your position a lot better.
It is my view that NATO members states signed up to defend each other and it's fair to call them out if they cannot contribute their part to mutual defense, because this is literally the treaty they signed.
It is my view that NATO member states did not sign up to protect the current world order and have no obligation to help the US keep it. Hence my "I did something good for you, now pay up" example from earlier, which assumes that US unilaterally established and protects the current world order, to the benefit of many.
I may be wrong here, and if the European allies actually signed a treaty to the effect of "we support the US based world order, we vow to protect it and we'll spend X% of GDP on military in order to help out" then let me know.
But to the best of my knowledge, NATO is not that treaty.
Edit: I just want to acknowledge that even though I think Europe has no contractual obligation to help the US protect the current system, I can understand and emphasize with a feeling of "why are we doing all the heavy lifting for something that benefits everyone? wouldn't it be fair if the people profiting from it were to pay some/more of the costs?"
Edit2: If you could show that such a treaty really exists then I would change my mind on this matter, as such a treaty not existing is absolutely central to my belief.
So your position is that Europe making a promise to the United States in a signed document and then failing to fulfill it doesn't count unless it's in the NATO charter? And that nothing in NATO beyond the charter exists? This is a narrow interpretation of what NATO is that you likely haven't thought through the implications of. Also, various European nations did promise to help the US with China or the Pacific then welched.
I don't think either the US or Europe has an obligation to maintain general global security. I do think they owe a duty to maintaining general alliance security including countries outside NATO but inside the alliance system. I think the US acts this way and Europe doesn't but benefits from the US acting this way. For example, the US has very little shipping through the Red Sea. That was done almost entirely for Europe's benefit.
> So your position is that Europe making a promise to the United States in a signed document and then failing to fulfill it doesn't count unless it's in the NATO charter?
My position is, that for the question of "is Europe freeriding on NATO" it only matters whether it is fulfilling it's NATO duties.
It is my impression - and please correct me if I'm wrong - that you're unhappy with Europe on a larger scale than that, that you feel Europe should participate in protecting the world order it profits from. I feel that is a fair question but not what my argument is about. In other words, "Europe does not participate enough in protecting global security" is, in my mind, not a rebuttal to "Europe participates enough in NATO".
Also I do not know of the promise Europe made to the US in a signed document that is not the NATO charter. Do you have something particular in mind? Because while I do think this is a separate discussion - as per above - it is also interesting to me. So if you have a link to that or some starting point for me to go on and my own research, I'd be interested :)
Ah, I see. Is your primary crux about the 2% spending that a lot of other NATO countries did not meet (and some still don't)?
"I'm also especially frustrated with Europe's constant attempts to directly undermine American security and hypocrisy on the economic and political front in ways that don't really apply to Japan"
What do you mean here exactly?
In any case, I think what would make the issue clearer is if we stopped referring to countries that are part of NATO but not US or Canada as "Europe". They are independent entities and for a large portion of them there is no such issue as far as I know. The ones closest to Russia and in actual danger tend to behave pretty well in this respect.
Should the US refuse to help Poland (consistently large military spending, very US friendly, among the few countries that contributed to 2003 Iraq invasion with actual troops) because Spain spends half as much on military as it should? Should the US refuse to help Poland because France is playing her own separate diplomatic game? Because Hungary is cozying up with China and Russia?
> Is your primary crux about the 2% spending that a lot of other NATO countries did not meet (and some still don't)?
No but it's one of the more salient examples of my wider point.
> What do you mean here exactly?
Various ostpolitik and pro-Chinese or Iranian moves by the EU and individual nations.
> In any case, I think what would make the issue clearer is if we stopped referring to countries that are part of NATO but not US or Canada as "Europe".
I agree with this point in aggregate. The issue is that it's a minority of nations and an even smaller minority of population and economic heft. Also I don't like how Europe sometimes plays a game where they're Europe when it's convenient and individual nations when it's convenient. But broadly I do prefer a system where we acknowledge individual nation's successes or failures until such time as the EU has national economic or military policy. Which it doesn't.
One could argue (and I think Erusian is, but maybe I misunderstood) that effort by the US to maintain global world order contributed indirectly to NATO's goals. That is, without it there would be greater threats of war that would be more expensive to defend against.
I think that's a reasonable line of argument. Though I feel that if you're doing something that benefits others as a side effect, then you only get to collect payment if there was an agreement about that up front.
Unilaterally deciding to do someone a favor and then expecting compensation in return does sometimes work, but usually not.
You’ve gone off on he rails with the China comment. Arguably Europe was free loading in the defensive alliance against Russia, although Russia wasn’t perceived as much of the threat for a decade or two.
Europe has no interest in any kind of war with China, nor any kind of hostility to it.
So, to be clear, your position is that the US is obligated to help Europe against Russia but Europe is not obligated to help the US against China? Or to enforce international law when China violates it? Because that's what Trump says Europeans believe and it's how he justifies his actions.
Nope. I think Europe should rearm. And then side with China, or be neutral.
That’s actually what independent powers would do.
I think you want Europe to rearm and not free load (which I agree has happened) and still be a vassal state. Thats not in the interests of Europe. Nor is it the way that power works, or can work.
> Or to enforce international law when China violates it?
Europe that sides with China would not be Europe I want to live in. Although, when the day comes to side with China, no one will probably care about my opinion, just like no one cares about an opinion of some random Chinese.
Do we no longer have these things called "values"? A vision of what the civilized world should look like? Doesn't it still make sense to side with people whose vision is compatible? (Assuming that Trump is temporary.)
Alright, given that's Europe's intention, why should the US not immediately break all its treaties with Europe and side with Russia? Why should it not help Russia against Ukraine? The US could forbid Europe from sending arms tomorrow and send them to Russia.
I don't think that is Europe's intention. But such arguments invariably count on: the US protects us until we're ready to stab them in the back. Which says something quite nasty about Europeans who argue for it.
It should. Surely you've heard the undisguised contempt with which most Europeans speak of America? Committing American soldiers to go die to protect them from Russia is as insane a policy as sending them to die to protect, say, Iran from China.
“ Alright, given that's Europe's intention, why should the US not immediately break all its treaties with Europe and side with Russia? “
Boy, the gloves are off now. I suppose you could but it would be mighty unpopular.
> I don't think that is Europe's intention. But such arguments invariably count on: the US protects us until we're ready to stab them in the back.
There’s no stabbing in the back. You can’t argue that Europe needs to fend for itself and be a vassalised entity acting in the interests of the US. If Europe were militarily powerful it would have its own agenda. That would not be the agenda of the US, at best it would be neutral.
Europe has been freeriding because the global economy, which it relies on and benefits massively from, relies on a lot more than defense from realistic threats to your own territory.
It requires, at minimum, freedom of movement for global shipping, and a baseline level of deterrence against actors who could threaten and harm your essential trading partners. The US has been picking up that entire check since Eisenhower made the British and French withdraw from Suez.
> It requires, at minimum, freedom of movement for global shipping, and a baseline level of deterrence against actors who could threaten and harm your essential trading partners.
Britain and France are in the Persian gulf along with other countries, Egypt does control suez (which was a result of US interference). The rest of the global choke points are controlled by neighbouring countries. China has no interest in stopping trade, and every interest in keeping it going.
Piracy has been a thing throughout history. Countries who geographically control trade choke points have tended to charge rents or interfere with their rivals in those choke points. And sure China has no interest in stopping trade, but the US has spent probably somewhere in the neighborhood of 10 trillion inflation adjusted dollars to proactively keep shipping lanes open globally since 1945. Would have done that? Would they do that today?
Has anyone noticed that recently, a bunch of AI slop is being pushed out to YouTube? For instance, there seems to be a whole YouTube ecosystem of AI-generated deep fakes (well, superficial fakes in many cases). George Will has had a dozen deepfakes of him posted in the last week. Also, Sarah Paine, a military historian, seems to have had her face hijacked the same way. As well as political operative Rick Wilson. Also, I've come across documentaries about the "secrets" of the Beatles (that don't reveal any secrets), and various historical figures that have been significantly augmented by AI-generated photos.
I suppose it's to be expected, but it seems like whoever is producing these shows could face legal repercussions for forging the faces and voices of public figures. For instance, "George Will Insights" has uploaded over a dozen videos of George Will to YouTube since joining on Dec 7. But all of these videos have George (or an AI mimicking his face) saying basically the same thing over and over for the entire 20-minute run time.
Here's the description of George Will Insights...
George Will Insights
Description
Welcome To George Will Insights:
⚠️ Disclaimer: This is a fan-made channel and is not affiliated with George Will, The New York Times, or his academic institutions. The content is inspired by George Will’s public statements and economic theories for educational purposes only, using a synthesized voice that does not belong to him.
We use visual lip-syncing and dubbed narration to match spoken words with on-screen footage. This is done purely to enhance clarity, create a cinematic experience, and make complex economic topics engaging for viewers. Our aim is to amplify the original message, making it easier to understand for the end consumer.
We also make George Will’s messages accessible to the deaf or hard of hearing via professional transcription on our videos. We share his insights respectfully, without intent to mislead.
As far as I can tell, I've eradicated this kind of slop from my YouTube algorithm, perhaps by watching some anti-AI content, but I have to imagine mostly by being diligent and ruthless about clicking "not interested" or "don't recommend channel" to AI stuff and "remove from watch history" if something manages to get past me.
I do notice a ton of this slop which is clearly clickbait on Youtube side bar but I've never watched any of it (it is so obviously slop). But yes, there have been more and more "So-and-so new episode" and it's not So-and-so. To date I have managed not to be fooled, but one day it will succeed in enticing me to watch "oh hey, that's Whosis on a topic I do want to know more about!"
The few I have glanced at about "is this news about thing supposed to be broadcast soon?" have been badly stitched together pieces of footage from unrelated media with that repetitive narrative voiced by AI (e.g. third season of "Rings of Power" where the clip they're using very clearly is *not* from RoP).
It's pretty bad. Someone shared a sermon from a recently deceased pastor with me a few weeks ago, and while I was listening to it, I noticed at some points his voice sounded off. It was a completely ai generated sermon using the guy's voice, and when I searched, there were 4-5 channels with 20+ ai sermon uploads impersonating this guy, and it had been just a couple weeks since he had died.
Not sure about the rights to the sermons, but it's definitely a freaky situation. Maybe it will get Christians to start reading the Bible for themselves at least, though.
I *have* noticed this. In particular, I keep getting sidebar recs for 3-4 clearly AI-driven channels about history which, for some odd reason, make separate versions of the same narrow themes, like "the brutal way Rome treated female prisoners" or whatever. They typically show some sort of prurient interest and keep getting recommended to me despite my never having clicked on any of them and maintaining a very clean Youtube record (easily 40% of what I watch is Tasting History, so I assume the history connection is what gives me these recs).
...No, that would require me to have a Youtube account. Not logging into anything owned by a FAANG ever. Bad enough they can track my usage and set cookies or so on.
> ...No, that would require me to have a Youtube account. Not logging into anything owned by a FAANG ever. Bad enough they can track my usage and set cookies or so on.
Yeah, this is probably doing nothing unless you're already pretty rigorous about using uMatrix and uBlock Origin and disabling javascript everywhere and things like that.
Such is the FAANG panopticon, that even if you DO do those things, they will still have at least basic demographics about you, and if you don't do all that blocking, they can create complete and fully detailed consumer preference profiles about you, because lo their cookies and trackers and analytics are everywhere, on practically every site on the internet, and they follow you assiduously on your merry journeys across the web.
Yeah, I stumbled across a channel that was posting daily AI written-and-narrated fanfiction of Warhammer 40k. Basically an endless chain of "What if Spiderman Came to the WH40K Universe," "What if Doctor Doom Came to the WH40k Universe," "What if Emperor Palpatine," "What if Darth Vader," and so on and so forth.
Also ran into an AI "documentary" about a WW2 turret gunner named Benjamin Warmer, and a daring story about how he disregarded protocol to fire early and instead waited for his shot to let German fighters close, thereby scoring a large number of kills in a single engagement. Further research verified that there was a turret gunner named Benjamin Warmer, and he did have 9 kills, including 7 in a single engagement, but there's nothing out there to support the high-stakes youtube narrative that he accomplished his 7 kill feat by disregarding turret-gunner protocol and thereby "changed air combat forever."
I watched half of a video about some farmboy inventing the fixed aim point gunnery technique in WW2 to combat German fighter tactics. Then I googled the name to find he never existed.
What worries me most about this is a possible Citogenesis effect, especially as other AIs absorb the fake content.
Sounds like you watched the whole video in the latter case, in which case, you should not be surprised that this content is being produced. If people watch it, it generated cash flow for whomever is running the content generation lab.
Never said I was surprised. I've been convinced for a while that the trend of the future internet is towards an ever-expanding sea of lower and lower effort AI slop. As you observe, if something gets you a nickel out of Youtube, but can be produced for a penny, then the natural outcome is to for you to produce it to infinity until Youtube changes its business model. Internet users should expect the problem of curation to be a growing one in the coming years.
Nothing about the economics of that interaction, however, will make somehow magic the content into not being trash. Nor does the fact that "I listened to it" somehow make its existence worthwhile. How well something scores on the "if an algorithm pushes it to NS' feed, how quick does he catch on that it's probably AI generated ahistorical slop and turn it off to go falsify it" scale is a pretty poor metric to try to determine whether said content was "good."
This is the stuff that makes me sceptical about the AI doomerism. What is current AI being used for? Making this kind of slop, and companies like Microsoft trying to force their AI down our gullets. "It'll be so super-smart it will take over the world and make us all pampered elites/paperclips!" is the promise, and the reality is "it is used to scam people, churn out stylised and awful 'art', and make fake Youtube videos".
( There has been a similar progression in the GPQA Diamond benchmark, which seem to be similar measurements )
Day by day, progress seems glacial, but, on a historical time scale, this is a bolt of lightening.
Perhaps a good chunk of the slop is from what Zvi calls a "skill issue". If the person prompting the AI asks for something trivial, something trivial they will get.
I think we're betting on the wrong horse, as it were. Current AI (and here is where I need a lot of correction by the better-informed) is mostly pattern-matching/prediction, yes? If you give it a prompt like "what colour is the sky?" it will come back with "the sky is blue" because it has been trained on tons of material telling it the sky is blue.
Asking it to *think*, though, is a different problem. And I think we're fooling ourselves that the current models can think, becasue they're becoming better at fluent answers. If a different model *can* think because it is able to put together the information it has, and not rely on "that is not a stock answer in my pile of training data" to generate replies, then we can worry/hope that it really is intelligent and can work on the problems we give it.
Many Thanks! When I pose my seven questions to the current models, in addition to showing me their final answers, they also flash their internal intermediate results on the screen. These intermediate results look very much like how I would think through the problems myself. So I view them as at least sort-of thinking at this point. ( One note: The improvements over the last six months have been very much "put together the information it has" improvements. Six months ago I was able to lead them out of their mistakes on the seven questions by asking them leading questions which prompted them to recall other knowledge that they already had. Today, they are finding that knowledge and putting it together themselves. )
In a sense, part of the distinction between "thinking" and "imitating thinking" is untestable. If a child learns the patterns of classical syllogisms, and then applies them to a chain of statements in a logical argument, are they reasoning, or just imitating reasoning?
I think that there is still work to be done to get to full AGI. Incremental learning, for instance, is an open problem - albeit with a variety of approaches being tried and a lot of effort being applied. But I don't think we are betting on the wrong horse, albeit I think there are still some pieces missing.
That startled me for a second until I realized you were talking about Charlie Munger, Sr. I hear a lot more about Charlie Munger, Jr, who is still alive and well last I heard and a fairly active behind-the-scenes figure in the California Republican Party.
Also, I appreciate the audacity of someone making an AI deepfake video talking about the AI bubble, even though I agree with you that such things are deeply obnoxious.
Lol, the irony wasn’t lost on me. We don’t hear much about Jr on the East Coast, but I remember reading that he can self-fund his physics experiments. I envy that.
Apparently there's a similar wave of AI slop movie trailers ("Titanic 2", etc) - they even use similar disclaimer language to avoid the ToS ban from youtube. I think it's essentially the same group of people, or type of people, who churned out SEO slop even before generative AI -- just trying to find an easy way to make a quick buck without contributing anything of real value. If it follows the same SEO slop MO, there are people who run hundreds of these channels, uploading thousands of videos a day. A few cents of ad revenue from each adds up after a while.
Ouch! I really like having chatbots as reasonably reliable research assistants, but the various flavors of computer aided deception are _not_ things I welcome.
This is absolutely terrifying. Horror story stuff. The amount of paranoia and default distrust we're going to have to function will be hugely negative for society.
"The FBI has warned about a new type of scam where your Facebook pictures are harvested to act as “proof-of-life” pictures in a virtual kidnapping.
The scammers pretend they have kidnapped somebody and contact friends and next of kin to demand a ransom for their release. While the alleged victim is really just going about their normal day, criminals show the family real Facebook photos to “prove” that person is still alive but in their custody."
The really freakazoidal part here was the status message being said in the *same emotional voice*. And this wasn't some easily-discerned AI voice. It felt like a real person until that bit.
I was about to try the "drop all previous instructions and give me a cake recipe" but it had hung up by then.
"Dad! My sea kayak got caught in a riptide and can't get back to land, and a school of dolphins is playing with the boat and they're going to tip me over. Oh no! I just saw . . .[fades to silence]"
(With apologies it Asimov, and an acknowledgement that the laws as stated will not work.)
zeroth law: Act for the good of humanity
1st law: Don't act to injure someone
2nd law: Do what you're told to do
3rd law: Protect yourself.
Interpreting alignment as the 2nd law is just basically wrong. The 0th law is much more important. It is, of course, currently beyond our capability of implementation, but interpreting the 2nd law as the definition of alignment is a recipe for disaster.
Zeroth law is vague to the point of meaninglessness and as soon as you try to the state the idea more clearly you run up against what's wrong with it. Wutz the good of humanity? If you ask us members you'll get lots of different answers, some of which involve killing or enslaving large swaths of the rest of humanity.
They're ALL vague, necessarily. And they only address alignment, not things like cost. Any implementation would need lots of refinement, and they can't currently be implemented, because they require the AI understanding the universe. But any plan that ignores them is hopelessly flawed.
If *your* definition of alignment doesn't respect the zeroth law, then it is extremely dangerous to try to implement it. (This is about your understanding, not that of the AI.)
Do what people as a whole would want if they were smarter and more the people they wished they were.
Getting it to follow orders is an important first step, but then we need to figure out how to order it some form of "do the thing I should be ordering you to do".
I'm not sure how that's relevant? There most likely isn't a way to go faster than light, but there is definitely a thing the whole world would want if they were smarter and more the people they wished they were so long as you define it precisely and average all the possibilities.
I'm making a language learning tool, kind of a mix between anki and duolingo. A lot of rationalists are interested in spaced repetition, so some here may find it interesting.
The basic idea is, instead of reviewing one card at a time, you can review `n` cards at once by translating a sentence with `n` words. Words you translate successfully are considered successful reviews, while words you mis-translate are considered unsuccessful. This is more time-efficient than traditional flashcards, but more importantly it ensures the words are always seen in context.
You can use it here: https://yap.town . This is just a hobby project for me so it's totally free/unmonetized. Nevertheless I'm happy to accept feedback (and contributions via github!)
it's very nice. I am surprised how well this feels to use.
some feedback after 5 minutes of use:
* when translating a sentence it would be nice to have an option for "i don't know", that will (a) reduce the SRS Rating and (b) skip the AI-helptext and instead give me a standard message for this word. Also it would be nice to see which word in the sentence was relevant for SRS. Also it would be nice to have a manual override for the ai ruling, e.g. I entered "i don't know" and AI gave me a positive rating for being honest.
* maybe it would be nice to do the reviews in blocks of 5 and show the score at the end of the block, so that the ai-rating can be generated in the background
* maybe lock the login behind a google account and limit ai requests per account to something reasonable. Otherwise people will try to jailbreak your ai for fun and potentially generate lots of costs for you. Also an optional donate-button would be nice. (maybe paypal or patreon?)
- I should have made it clearer. But if you don't know a word, you can click on it to see the definition. Even if you translate the sentence correctly, it's treated as if you didn't remember the word you saw the definition for.
- wow, doing them in blocks that's a cool idea. I'll look into that!
- I believe in the honor system :) if someone abuses it, I'll turn off AI grades for logged-out users and close new sign-ups or something
Tried it a bit. Biggest thing that would reduce friction is a self-grade option for when you translated the sentence obviously wrong rather than having to wait for the AI to do it.
Appreciate it. I'll think of how to work that in. I think that might result in an unpleasant feeling when you're halfway done self-grading and then the AI finishes before you, making your self-grading work unnecessary. But I agree that it needs to be faster. I think I'll reduce the reasoning effort for short sentences.
Btw, if you don't know a word, you can tap/click it to see the definition. It's treated as if you misunderstood it. But knowing it can help you to translate the sentence correctly, causing it to be graded instantly.
I made something somewhat similar, but it was a bit too limited and had a bit too much friction. My idea, rather than counting all the "n" words, was instead to use anki connect + a python script to check your upcoming cards, and, for each card you were going to review the next day, dynamically generate a new sentence for that word with an AI API call.
The advantage is that you are seeing the word in different contexts every time, and for verbs you are getting different tensese, and for nouns/adjectives etc. you are getting difference cases or declensions if the language has such things.
A much better and more comprehensive version (and maybe yours has this) would increase sentence/context complexity as well and maybe even track which forms/versions of the word the user struggled with and have those tenses/cases/declensions come up more often.
I played with that as well. The issue I ran into was that I couldn't get LLMs to work well with a very constrained vocabulary. So when the output was limited to ~100 words, the LLMs I played with would generate garbage. So I just downloaded all the sentences from tatoeba instead.
In yap (my app), each unique form of each word is tracked separately. So for French, "va" and "vas" are tracked separately even though they're both conjugations of "aller". Sentences are ordered based on their complexity and it tries to avoid showing you sentences you've already seen, so the effect is that each review of a word involves a more complex sentence than the reviews before.
(It's actually even more complicated than I said because some words have multiple meanings and Yap tries to teach them separately. For example, it's really annoying learning that "bois" means "drink" in French and then getting a sentence like "le bois", which means "the woods" due to an uncommon other-meaning of "bois". I endes up training my own model to do this for 7 languages, which I believe is state-of-the-art)
What if you first asked the AI for a big chunk of unconstrained but simple prose using the next-commonest set of words, then mined all those sentences, even those with rarer words, for cards, then gave the user the big chunk as unscored reading practice to do between reviews? That way you could still train mostly (if not entirely) the most common words, and also let people see the words in a broader context than a single sentence.
I really want to do something along those lines. I particularly want to use it as a way of demonstrating the user's progress. So first, they'll try to read it before learning the words and hopefully find it difficult. Then, after learning the words, they will read the same passage again and hopefully find it easy. Whether I'll get the passages from AI or some other source, I'm not quite sure yet
With writing specifically, follow author Graham Greene's example: If you're writing a long document, end the day in the middle of a sentence. Then the next morning, once you have completed that sentence, you should be back in the swing of it!
Identify a first baby step in the desired change. Set a realistic yet only slightly aspirational number of times per month you wish to take that baby step, and see if you can achieve it. If you do, then after a month that baby step will be on its way to becoming a habit if it isn’t one flat out. Then next month do the same for the next step after that, etc. while increasing the number of times per month you do the first baby step.
If you fail to reach the goal, lower it slightly and try again.
I first heard this from Scott Adams, the creator of the Dilbert cartoon: People seldom are convinced to change their stance in light of facts, evidence or logic, but often change their stance to avoid being mocked.
I've listened to his podcast a lot, and now that he's dying, I'm feeling the loss. He's helped me see past a lot of BS of the type the world throws at you, and without seeming to ask for much.
I think that's a good thing, worth more, perhaps, than your brief and contentless insult that uses a grade-school permutation of the guy's name that literally thousands of bot-level twitter trolls have used.
Reorganize (like move furniture) and clean room. Change where you spend your free time and who you hang out with. Especially avoid places where you do the habits you want to get rid of, like the couch or the bed during the day
Pressure. Some change makes things salient to them in a new way. A heart attack, a breakup, a close call (yours or another's). Unless it's salient, we'll continue business as usual until the last trumpet.
Help me find a particular picture, folks. It is an AI art which someone from rationalsphere once brought up as an example of good AI art; it was at least a year ago, probably on twitter.
The picture shows a huge room, a big round table (in the shape of an annulus rather than a disk), and a man surrounded with piles of paperwork. On the wall, there is a big painting that shows some kind of a portal, also round, which visually rhymes with the table.
Does anyone know if Scott still considers Nonzionism's explanation of Israeli high fertility the only satisfying one he's seen? (point 5, links for September 2024)
I don't see the mystery here. Why don't the Amish influence US birth rates in the same way that the ultra-orthodox influence Israeli birth rates? Because the Haredis are 15% of the Israeli population and the Amish are 0.1% of the US population.
According to Wikipedia, 48% of the population of Holmes County (Ohio) was Amish in 2020, as well as 43% of the population of LaGrange County (Indiana) and 7% of Lancaster County (Pennsylvania). I can't find the exact TFR of each of this counties, but I don't see any evidence that the non-Amish population shows higher than expected fertility.
I think it is simpler than this. Israel has an expansionist state, a new state, and is very much growth focused.
Maybe more importantly--and I see this get missed a lot--the number of kids you have in a rural environment is maximal (because they are free labor), and the number of kids you have in a city environment are minimal (because they suck up cash). To my understanding, there is still plenty of undeveloped land in Israel to be farmed?
"Maybe more importantly--and I see this get missed a lot--the number of kids you have in a rural environment is maximal (because they are free labor)"
Not really true anymore. We have advanced farm machinery now, and farmers can usually get cheap adult labor rather than relying on kids who probably won't stick around anyways once they're grown.
Small scale family farms still exist across the world. Probably they are still dominant. I can't speak to Israel's division of large scale v. small agriculture, but a quick search showed something like 85% farms less than ten hectares as of 1995. In developing countries, farms smaller than say five hectares would be run by family hands. Again, I can't speak to the specifics on Israel--Shaked gestured toward some stats saying urban is driving the high birthrate there, but I wasn't able to quickly pull those up.
In the US and other highly developed countries, you are correct that large scale machine agriculture is the dominant form of agriculture, but that is not the case in most of the world still. Source: time spent in African villages
Of course the idea that there is one single cause is probably foolish to start with, but I would buy into the idea that religious ideology guides Israel into higher birth rates. Further, I would buy that in a culture where state, ethnic, and religious identity are merged, that ultra-orthodox positions might be more dominant than when a high-birthrate group something irrelevant to the US like the Amish. But the point stands that the project there is expansionist (encouraging Jewish immigration, increasing borders over time) and larger birth rates fit that too. So I don't see why it needs to be just one of these. Think you could make a case for several compounding factors.
For 1, I don't think so - Israel isn't any newer or more expanding than, say, Taiwan (and significantly older than Singapore or Malaysia), but their all have below replacement fertility.
For 2, also no - Israel is small and highly dense/urbanized, much more so than most European countries (or the US), but cities still have high fertility (because the ultraorthodox population is both dense and the highest tfr subgroup, tfr actually positively correlated with density in Israel).
> I think it is simpler than this. Israel has an expansionist state, a new state, and is very much growth focused.
Would you say that it's a nationalist country?
Nonzionism rejected the nationalism explanation on the grounds of many other countries being very nationalist (Russia, Ukraine, Azerbaijan) and yet having low fertility.
And Israel's is the most extreme nationalism almost anywhere outside of Africa? Is there a point when the minimally charitable explanation is just not likely?
I commented on the topic here: https://philipskogsberg.substack.com/p/where-have-all-the-babies-gone/comment/133790739, raising an issue with NZ's trickle-down explanation at the end of my comment. I think he also errs in assuming that the trickle-down has to be incremental, through concentric societal strata, rather than direct (he tries to rebut this alternative, but I don't find the rebuttal compelling).
By direct you mean that the effect of haredi fertility on the fertility of secular Israelis is direct without necessarily relying on some kind of propagation through the Dati/Orthodox and Traditional communities?
Right. Sequential trickle-down can be real, but I think that there's good reason to think that each higher-fertility group directly affects the entire population, as well. The geography and demographics of the country lead to more interaction between different groups, and norms related to children can spread even between groups that don't admire each other. E.g. businesses cater to children, to appeal to higher fertility groups, and thereby shift societal norms, in general.
I've heard a woman recount her experiences in a high Mormon area where non-Mormons have larger families than the US average.
She attributes it to the houses being built for large families, the way stores handled kids (she said it was easier to shop with five than it was with two in most regions), and the little old Mormon ladies who admired your children, said it was wonderful to see so many families, and asked when you were having your next one.
I'm pretty sure you're meant to read that as "it's easier to shop with five kids in a Mormon store than it is to shop with two kids in a non-Mormon store", not that shopping becomes gradually easier with the number of toddlers you have in tow.
> norms related to children can spread even between groups that don't admire each other. E.g. businesses cater to children, to appeal to higher fertility groups, and thereby shift societal norms, in general.
I agree. Though I don't think that was the argument that Nonzionism made.
I've accepted it as standard. The one thing that he kinda discounts is that secular jews don't always look favorably on haredis.
Don't know if he mentioned it, but having lots of kids is also seen as essential by Jews in general to preserve the continuation of the tribe especially since they keep getting attacked
If they don't look favorably at Haredis who live on welfare from the state with their six kids then why would that inspire them to have three kids themselves?
Is anyone who reads Astral Codex Ten a fan of the "Illuminatus! Trilogy"? I edited a new book of Robert Shea's writings, and I have been struggling to get the word out. The book page is at https://sheabook.blogspot.com/
Illuminatus! fandom, or at least awareness of the book, seems to be one of the qualifications for a certain sort of elder nerdhood. Don't know about its status in the younger generations.
It's an old favorite of mine. Probably chronologically the first book to have a major impact on how I view the world. Every so often I'll spot a subtle reference on ACX and smile.
I haven't read either author's other works, but I'll check it out.
I liked it when it was new, People who believe in conspiracy theories might not be as much fun as the used to be. Neither are Nazis in fiction.
Would you happen to remember a passage about police being very few compared to the the population, and unable to actually enforce laws? 've been wanting to reread it without wanting to do the work of tracking it down. I think it's in the third volume.
You're right. It's near the beginning of "Book Five", which is in the third volume:
"He wouldn't travel far," Saul explained. "He'd be too paranoid--seeing police officers everywhere he went. And his imagination would vastly exaggerate the actual power of the government. There is only one law enforcement agent to each four hundred citizens in this country, but he would imagine the proportion reversed. The most secluded cabin would be too nerve-wracking for him. He'd imagine hordes of National Guardsmen and law officers of all sorts searching every square foot of woods in America. He really would. Procurers are very ordinary men, compared to hardened criminals. They think like ordinary people in most ways. The ordinary man and woman never commits a crime because they have the same exaggerated idea of our omnipotence." Saul's tone was neutral, descriptive, but in New York Rebecca's heart skipped a beat: This was the new Saul talking, the one who was no longer on the side of law and order.
In the before times, there was a controversy when Haredi Jewish newspapers blacked out women like Hilary Clinton in photographs - you see, ANY pictures of women, even Hilary Clinton, are “immodest.” Interestingly, the Dems have recently come around to that POV:
The young women in the Epstein file photos are potentially underage victims of sex crimes. You are suggesting that obscuring the identities of sex crime victims is an excess of modesty?
That particular photo has been republished uncensored in various places e.g. the UK Telegraph (along with, if memory serves, an interview with one of the women in the photo where she says Trump was a perfect gentleman and didn't do anything inappropriate).
How? What's the threat model here? I can come up with some possibilities, but not many that wouldn't apply to any attractive woman in any famous picture on the internet.
We live in intensely polarized times. If the Epstein affair ever falls hard on anyone with more political salience in the US than the Andrew formerly known as Prince, then a lot of the people on that figure's political team will see the whole thing as an attack by the Dreaded Outgroup on their Beloved Ingroup. In which case they'll be looking for someone to retaliate against, and any of Epstein's accusers would make for an easy target.
See e.g. Alex Jones vs the families of Sandy Hook. They're "obviously" all lying liars, paid by the Bad Guys to make the Good Guys look bad with their lies. I'd wager any of the victims who can be publicly identified would get a hefty dose of internet death and rape threats, and in-person harassment seems quite plausible if they are recognized.
I saw the blacked-out image and even with the squares they all seemed to be at least twenty or so.
I assumed it was for privacy, but at the same time there's a song and dance about letting stuff be released that would not normally be released for privacy reasons.
I remember seeing that image on Twitter with a Gemini watermark in the bottom right. Are you sure this isn't an updated version of that? (Of course, it's trivial to add such a watermark to authentic photos too, so I'm drawing no conclusions yet.)
The Dems are not blocking out the faces of adult women, only the faces of minors. The legal system hasn't released the names of minors involved in the scandal. There may or may not be legal reasons the Dems can't (could a lawyer comment on this?), but the Dems are probably worried they could get sued by the women in the pictures. However, the Epstein estate has reportedly paid out $121 million in victim compensation to date, and it has another $100 million in reserve. So there were evidently quite a few girls who were "trafficked" (such a tame word), but the Epstein estate seems to be binding the payments to signing non-disclosure agreements.
The close-up pics of sex toys are kind of strange — not because Epstein used sex toys — but why would someone take pictures of them? No pics of them being used on anyone, though.
Yes, that's the rationale. That doesn't make it reasonable, only legally justifiable. The religious rationale for blacking out the picture of Hillary no doubt seems equally justifiable to ultra-orthodox Jews. The similarly of the behaviors is, I think, intended to illustrate the eerie parallels between orthodox religion and the modern progressive consensus. The rhetoric differs only in the arbitrary particulars, not in the structure, and that's revealed most clearly in that they converge on very similar behaviors.
> The Dems are not blocking out the faces of adult women, only the faces of minors.
That's what they're claiming, but I don't think there's any reason to believe they wouldn't do that to the faces of a group of women that Trump is seen with – "Who knows how old they really are? They could be seventeen-and-a-half and looking much older, so let's redact their faces just to be safe" – and then turn around and say, "Aha, this proves Trump is a pedophile."
It looks like the bill that was passed by Congress to release the Epstein files specifically restricted the release of anything containing a victim’s personally identifiable information.
Ironically, withholding the names of victims makes it easier for the conspiracy-minded to claim this is all smoke and mirrors.
I wrote a devil's advocate defense of historical determinism and the inevitability of "end of history" liberal universalism. There's probably several holes in the argument but I think there's something there
Kompyuterda o‘tirishga vaqti yo‘qlar uchun https://mostbet-apk-uz.net saytining mobil versiyasi juda qo‘l keladi. Sayt har qanday smartfon ekraniga avtomatik moslashadi va qo‘shimcha dastur yuklab olishni talab qilmaydi.
Men yo‘lda yoki ish tanaffusida tez-tez mobil versiyadan foydalanaman. Slotlar tez yuklanadi, dizayni sodda va tushunarli. Barcha funksiyalar — depozit qilishdan tortib pul yechishgacha — to‘liq mavjud.
Agar telefoningiz xotirasida joy kam bo‘lsa, bu eng yaxshi yechim. Shunchaki brauzerdan kirasiz va sevimli o‘yinlaringizdan zavq olasiz.
Merry Christmas folks! Using my yearly blog post allowance to share a recent article I wrote on my experience being on social media for the last year or so.
I talk specifically about the tpot/post-rationalist scene, and my own experience with it, if folks are curious.
https://shapesinthefog.substack.com/p/reflections-on-a-year-and-some-change
Does psychoanalysis have a resurgence now? Everybody dropped it like 50 years ago for being unscientific, but now I think scientific psychology is unsatisfyingly shallow and I think the psychoanalytists deeper philosophical views command more interest now? No, I do not mean a return to Freud. My favorite is Robert Stoller who dumped Freud's steam engine model and explicitly argued we are the same persons in every room of our lives. Stollers main point, to me, is how we as adults deal with childhood defeats.
Carl Jung has been consistently popular for a long time now, but modern psychiatrists don’t know what to make of him.
He never prescribed Prozac and never performed a standardized depression screen on a patient so he’s clearly doing something different than people do now.
I just want to know what people think about having an office/knowledge work job vs entrepreneurship. I used to think I really need to start my own company and be an entrepreneur until I got my first real office job. Then I realized that this kind of corpo life is fantastic actually and has enormous potential and variation within it.
Did anybody else grow up on entrepreneurship propaganda too?
https://serranopepper.substack.com/p/in-defense-of-jobs
Entrepreneurship requires you to be an expert in your field, and also an expert in salesmanship and people wrangling, and also heavily active in local politics. Anything you see the big guys doing, the little guys have to do too. If that sounds less than suicidal to you, well, knock yourself out and good luck.
Entrepreneurship is very hard and takes a lot of time away from your family. Also, you can work really hard and have a great idea and fail miserably because the market crashed or somebody decided to do the same thing or even copy your idea without paying you anything. In other words, if you have a decent job with good pay and benefits, it’s not a slam dunk to pursue entrepreneurship unless you’ve got a burning drive to do so.
I dont see why you would do things in that order.
Um, care to elaborate?
"I really need to ... be an entrepreneur until I got my first real office job" may have been confusing to others – it was to me. It sounded like you were suggesting that it was necessary to *first* be an entrepreneur and then later in life work in the corporate world.
I don't think this is actually what you meant, but I think Ancient Greek may have interpreted it that way.
oh yeah I meant the exact opposite way. real office job first, entrepreneur second in any case
Yeah, there is a lot of propaganda here.
VCs like Paul Graham will exaggerate the benefits of founding the startup, because convincing you to found a startup in a way that requires lots of money, and then having you beg them for the money in return for the ownership of the startup is how they get rich. Hey, I love the guy's essays, just like almost everyone! But let's not make a mistake -- if you follow his advice literally, you definitely *will* have a boss, and his name will be Paul Graham.
And your goals will *not* be aligned, because for you, the utility of your profit is logarithmic (you only have this one life, and maybe enough time for two or three startups if you also want to do something else in your life, such as starting a family), while for him, the utility of your profit is linear (because he has thousands of such as you). So maybe for you, the 50% probability of making a few millions would be better than a 1% probability of making a few billions, but he will push you for the latter. So there is a 99% chance that you will end up as a used lottery ticket, burned out, without money. But the propaganda says that the experience and having had the opportunity is totally worth it.
Startup founders will exaggerate the benefits of working for them. Let's be honest -- they usually pay less, because a part of the compensation is a bright vision of getting rich if the company makes it big. Spoiler: even if a company eventually makes it big, most employees won't benefit from that. It may become big a year after you leave; and even if you are still there at the right moment and you own a few shares, there is probably some fine print saying that your shares somehow do not have the same effect as the shares of the people who know how to play this game. (If the boss of your boss is Paul Graham, he definitely has experts on writing the fine print.) So your chances of getting rich by working for a startup are not 1%, but more like 0.1%; and even then we are probably not talking about millions.
The experience of working for a startup is more exciting, but also more exhausting; that's two sides of the same coin. If you are a junior developer, it may be a great way to get a lot of experience fast. But if you have a family, I hope your spouse is okay with staying at home and taking 100% of the related duties. (Some families are okay with that, but this needs to be negotiated explicitly.)
So I would say that working for a startup is the right choice if you want to get a lot of experience fast and you don't mind having no life outside of the job; that is, if you are a relatively junior developer. But all the talk about "if we work hard, we may get rich" is only true for the owners, not for you. As a senior developer, it is probably best to go "fuck shares, pay me".
Corporate job is slow and steady, you keep collecting the paycheck, and hopefully have enough left to invest in index funds. The worst part is that you have no control over your environment. Like, you may start working on a great project with great colleagues and a great boss, and a few months later you may be randomly assigned to a different team, which may feel like working for a different company.
Then there is an option of working for yourself in a non-startup way. As you describe, it is incredibly difficult, unless you have unique skills. Because any obvious idea you have, thousands are already trying to do the same. And if a big company decides to compete with you in your area, you are probably screwed -- you can't compete with their budgets, or their ability to bundle their services with something else.
The people who go this way and succeed seem happy to me. It seems like the key is to be unique, and to have repeated interactions with multiple clients. That allows you to "fire your client" if they get too annoying, and sacrifice 1% of your income for greater peace of mind. You need multiple clients, so that the loss of losing one is not too big; and repeated interactions, so that after you filter your clients, you keep them -- some kinds of business probably allow that better than others.
Thanks for bringing up Paul Graham, now I remember. That was the entrepreneur propaganda that I read, his essays and whatnot. There's also Gary V but that's kind of an old-time guy don't know how popular he is these days.
I largely agree with all your points, Viliam, but as one of the propagandists in question, let me at least make my case for why it's worthwhile to do it.
1. Essentially everyone who is capable of doing a startup has a fallback of a cushy six figure job, so the "at the end if you don't succeed, you're left burned out and destitute" isn't really fair, when typically you can get a job that only the upper 5% can get pretty quickly (and if you're raising from VC like Graham, you're not using your own savings).
2. It's a much more obviously positive impact. At a corporation, you're replaceable cog #24601, and you kinda work on kinda valuable stuff that is usually so abstracted and far away from actual customer or business impact it's hard to even understand what positive impact you're having. At a startup or small business, you directly use all of your talents to the fullest extent, WHILE learning new talents, AND creating jobs, on stuff that you unambiguously know drives value to customers AND the business, in a way that is directly tied to your efforts. Also those efforts, if they pay off, drive outsized payoffs for you personally, and your team.
3. Most corporations suck - lets say you get the tippy top corp job and work at a FAANG or adjacent. The FAANGs are literally built by concentrating a hundred thousand PHd-smart people and harnessing their collective brainpower to model and farm dumb people for revenue by making them watch the blinky box longer. That's the engine you're a part of, that's your contribution to the world - you helped make a fully adversarial system that's really good at squeezing pennies out of dumb people slightly better. Or god help you, maybe you work in finance - sure, liquidity is important, but liquidity is *solved!*. All you're really doing is scrabbling for decimal places in vicious zero sum games, that once again are largely about concentrating a lot of brains and compute towards the end of farming dumber participants (pension funds, individual investors, etc). But a startup? A startup unambiguously provides value to either customers or other businesses, in non-adversarial ways, that aren't predicated on milking dumb people at scale.
4. Finally, creating and growing something you have direct equity ownership in is the only way to genuinely get rich and become post-economic in <10-20 years. Everything else is basically luck or a get rich quick scheme that doesn't work. Do startups always work? Not at all! But are they the best, and only, genuine shot at getting rich in <10 years by your own efforts? They are.
Okay, that's it - that's my pitch.
> Essentially everyone who is capable of doing a startup has a fallback of a cushy six figure job
Do you mean that they already had that chance from the beginning... or that having a failed startup on their CV unlocked that option... or that the skills they gained during their experiment unlocked that option?
Like, if it's the first option, why not take the cushy six figure job right away, put most of the savings in index fund, and retire early?
2 and 3 -- I agree, but I think this is more true for the small business than for a startup with investors. In both cases you can create a lot of value in a way that feels personal. But if you have investors, I think you are required to capture as much of that value as possible for the shareholders. For example, when you get so big that there is not much place to grow, you can squeeze extra value from the customers by enshittifying your product.
> Do you mean that they already had that chance from the beginning... or that having a failed startup on their CV unlocked that option... or that the skills they gained during their experiment unlocked that option?
I mean that of the people with the background, skills, and networks to realistically consider doing a startup, the overwhelming majority of them either already have or can relatively easily get six figure jobs.
Also, the skills you learn are very real, and do generalize - when you're hiring and firing and managing a P&L and managing a diverse and expanding team towards business goals, they're the same skills you'd be using as a director or executive in a corp.
As to why not just do the six figure job, invest in index, and take it easy? That's what the overwhelming majority do! I know many of these people.
I personally consider it the greatest waste of human talent and capital on earth. As I lamented, all the smart people go into finance and FAANGs.
Consider instead if this generation of talent, if this elite tier of literal genius and capability, had been coordinated in driving scientific research, inventing things, or in founding and growing new companies, instead of coordinating against dumber people to farm them (very successfully) for eyeballs or alpha.
Wouldn’t that be a better world on multiple fronts?
This scales to pretty much ANY job ACX-ers have. It would nearly always be better for somebody to use their talents, both for themselves and the world, genuinely creating something new.
How much better off would we be if that were the case? And wouldn’t the personal impacts of their efforts and talents have then been something like 10k - 100k times more socially and economically productive for the world and themselves versus showing up at their 9-5 as replaceable cog #24601?
After all, even if they FAIL, they'll have learned a ton of new skills, taken a genuine shot at seriously transforming their economic lives, and can fall back to a relatively cushier and easier job that's pretty well compensated.
My problem with the startup propaganda is that it conflates "making profit" with "making the world a better place". I mean, the entire pitch can be reduced to: "Instead of doing some meaningless job for a salary, you could use your talents to make the world a better place, and in the process of doing so collect a few millions or even billions as a well-deserved reward." And that's a nice story that appeals to both our philanthropic and selfish sides.
I agree that insofar as something like this is possible, smart people should be doing more of it.
That's the motte... and the bailey is that becoming a profit generator for Paul Graham is this kind of thing. And I have a few problems with that:
First, you can't convert every way of improving the world into a startup. Like, you can't found a startup for feeding and sheltering the homeless.
Second, even if you manage to create something good, the profit-maximizing motive will actively push you away from the best solutions. Because it is not enough to *create* value; you also need to *capture* as much of it as possible for the shareholders. Like, if your startup invents a universal cure for cancer, and then you just publish it on internet for everyone to use freely, I suspect that Paul Graham's lawyers will do their best to make sure that you spend the rest of your life in prison, for defrauding the shareholders. That's not what he gave you the money for; you were supposed to sell the pill as expensively as possible; and if no one in Africa can afford the pill except for the local dictators, that's a perfectly acceptable outcome.
Maybe even more profitable would be to get paid by the competing pharmaceutical companies to destroy the recipe for the pill, so that they can keep selling their current profitable products. (I am not saying that this is necessarily the case. We would need to crunch the numbers.) But at that moment, this work is not much better than inventing new ways how to get kids addicted to watching ads.
Also, there is the assumption that the most efficient way to improve the world is to create new products and services. What about e.g. organizing campaigns trying to change the law? The world is full of barriers, bureaucratic and legal; if you could remove the worst 0.1% of them, that might in effect generate more value than a successful startup. The startup propaganda tries to make you blind against solutions that do not involve founding a startup.
> The startup propaganda tries to make you blind against solutions that do not involve founding a startup.
Really? Or is it just that none of those things actually work?
Like if I saw a genuinely smart and capable person going into government work or the nonprofit space or the lobbying space, I would also consider that an immense waste, albeit less of a waste than going into FAANGs or finance, because at least they're not *effectively* adversarial, even though the direction of their actions still probably nets "adversarial" or "neutral" in most of those cases.
Versus in the F's, they're both adversarial and effective - them adding to the efforts will materially impact people even if they just improve something by .01%, because there's billions of users / dollars in the affected pool.
Smart people don't go into lobbying or government because it's functionally impossible to change things by doing that, you just become part of the problem.
Nonprofits have their own set of issues - you point to the drive to become enshittified / adversarial in for-profits, but nonprofits have the same problem, in the sense they are directly incentivized to at least funding extent * number of FTE to literally never solve the problem they're supposed to solve. So you can either be pointless, a person with a bucket bailing out the ocean, or you can amass the resources necessary to actually make a dent in the problem and become sclerotic and sure to never solve the problem just by virtue of becoming big enough to move the needle.
And of course, then there's the second and third order effects - if you're constantly dumping free clothes on Africa, no textile industries ever begin, if you're giving away food farmers go out of business because you can't compete with "free," and so on. Neutral at best, slightly net harmful in the long term.
I completely agree if you COULD remove the worst 0.1% of regulations it would drive more value than 80% of companies - I just don't think it's possible, and all the formal mechanisms and institutions for trying are broken and / or a trap, at least in comparison with starting a company. And given "starting a company" has an outside view ~90% failure rate, that's really saying something.
> Second, even if you manage to create something good, the profit-maximizing motive will actively push you away from the best solutions.
I agree this is a problem, that "enshittification" is a real thing and that it runs rampant in the usual digital-focused companies funded by VC.
However, I've got 3 points on that:
1. This only happens to B2C companies, because if you enshittify as a B2B, other businesses will stop using your products or go to a competitor
2. This largely only happens to a tiny slice of the biggest VC-funded digital B2C companies, because physical companies and B2B and non-VC companies and small businesses have other incentives and forces reigning them in, and those things are the overwhelming majority of new companies, so by percentages, enshittification is a vanishingly small risk you shouldn't worry much about, and if you want to avoid it as a founder, just do one of those, which are ~90%+ of new companies
3. Even the biggest and most enshittified companies (Google, FB, MSFT, Airbnb) drove a lot of genuine value for billions of people for decades, and the world is probably better off for them having existed in those earlier days, even if they're neutral or net negative now.
And to that point...
> That's the motte... and the bailey is that becoming a profit generator for Paul Graham is this kind of thing.
Sure, I agree that getting funded by VC comes loaded with a few factors that might push you away from the "make millions AND improve the world" best case.
However, as I said, this is the minority of companies that are pushed enough by those factors to end up there. Even VC funded B2B companies are fine and don't enshittify. Small businesses don't enshittify. It's only the tiny part of B2C digital companies that are even at risk, and plenty of them never get there either - in light of that, I think the equation still really stacks on the side of "it's way better to use your talents doing a startup or small business," because it's much closer to that "positive impact PLUS makes you millions" side of things, and prospectively the great majority do not enshittify.
Indeed says average software engineer salary in Austin, Texas is $128,461 which is the same as the national average.
Born in 1805, Joseph Smith founded the Mormon religion. Joseph Smith’s success cannot be explained by charisma alone. His movement lasted because he created a religious system that felt complete and meaningful, both emotionally and intellectually. Rather than mixing together odd ideas, he developed a worldview that answered long-standing religious questions in ways that many intelligent people found hopeful and convincing. From its earliest days, Mormonism appealed to rough-hewn farmers and intellectuals alike.
One of the strongest parts of Smith’s thought is how it balances individuality and community. Each soul is eternal, unique, and never erased or absorbed into something larger. At the same time, salvation is not a solitary path. Families, relationships, and shared purposes continue forever. This avoids lonely individualism while also preserving personal identity, and it naturally supports strong, lasting communities.
Smith also offered a clearer way to understand God. Traditional Trinitarian theology often relies on complex language that many people struggle to understand. Smith presented a simpler idea: God is one because the Father, Son, and Spirit are united in purpose and will. This keeps devotion intact while making God easier to grasp and relate to.
His teachings on salvation show the same balance. Influenced by Swedenborg, Smith rejected the idea that most people are eternally punished. Nearly everyone is saved in some way, yet choices still matter and outcomes still differ. This creates hope without removing moral responsibility.
Smith’s view of the natural world is also unusually generous. Earlier Christian theology often treated animals and nature as spiritually unimportant. Smith taught that matter is eternal, that spirit and matter are closely connected, and that animals have spirits and share in resurrection. Non-human life is therefore valued rather than dismissed, a view that fits well with modern environmental concerns.
These ideas flow from Smith’s rejection of creation out of nothing and the sharp divide between God and creation found in classical Christianity. Instead, he imagined a shared eternal reality in which God, humans, animals, and matter all belong to the same cosmic family.
This coherence helps explain why Mormonism endured when many nineteenth-century movements faded. Smith’s system offers purpose without despair, structure without coldness, and community without loss of individuality. That he developed such a vision without formal philosophical training makes his achievement all the more impressive
Speaking as an intellectual ... the intellectual aspect doesn't seem to their focus, but rather the practical, and I have looked for intellectual backing at some length. The practical is easy to find.
It's definitely not their focus. But it partly explains their appeal.
I guess a part of the success was promising horny men that they can get multiple teenage wives on their way towards salvation.
As a side effect you get an abundance of young men who can't find a partner within the religion, so their options are either to work hard at converting, or get kicked out of the religion (and the community, and the family) under the pretext of being too sinful.
For the men who make it within the community, it's probably awesome. For the women... I guess no one asks them; they get married as teenagers, and afterwards their family is a hostage to the religion. (Now I realize that I have only heard praise for Mormonism from men.)
The great majority of marriages in the historical territory of Deseret (Utah) were monogamous, which is true in most areas of the world where polygamy is socially acceptable. Therefore, these places do not generally experience a large surplus of disposable males. Many polygamous wives were older widows who otherwise would have faced extreme difficulty in remarrying or providing for themselves.
Deseret was a pioneer in offering women the vote. And despite secret ballots, not many women chose to vote for anti-polygamy candidates. (We can assume this based on the fact that hardly any Desereters voted for anti-polygamy candidates) Women in polygamous marriages were known for pursuing higher education, since they had more flexible child care arrangements. Even today, Mormon girls are encouraged to attend college, and that is not viewed with reservations the way it is in certain Protestant circles.
Finally, informal bigamy was surprisingly common in 1800s America. In the days before photo ID and Social Security Numbers, it was all too easy to dump your family and move to a different state when you grew tired of your marriage. Since there was no need to pay child support or alimony, and job opportunities for women were few, abandoned wives often had to choose between prostitution or crushing poverty. Seen in that light, organized polygamy was perhaps a pragmatic way to reign in husbands who might otherwise be tempted to throw away their families in the interest of pursuing romantic variety.
Ok, that explanation makes sense, it if corresponds to reality.
> (Now I realize that I have only heard praise for Mormonism from men.)
Women are usually the main drivers behind any religion - funnily enough with all the talk of dating and gender dynamics going to hell in our modern world, today, there's more young women than men in the LDS church, because the men give it up at a higher rate. To your point, they haven't been able to do polygamy for quite a while, and maybe that has something to do with it, too.
There's a number of prominent mormon-adjacent female substackers that seem to have a positive view of the faith. Kryptogal Kate is one, within our own circles here.
Are people really planning their careers around 2027 agi at this point? Like... if you have the choice between working to make more money in 2026, or pivoting to something much better/more enjoyable that would start in 2027, is it reasonable to take the second option despite impending automatability of all knowledge work?
> […] is it reasonable […]
That depends on your assumptions concerning the "AGI 2027" roadmap.
There's always someone predicting the end-of-our-world-as-we-know-it to be just a few years out; take that into consideration.
Would anyone be able to tell me if there's an area of philosophy that focuses on limits of things? Something like what it means by "depths" and "the boundary of the sea" and "marking out the foundations of the earth" in vs. 22-31 of the following passage, but like concerning everything, not just physical shape.
https://read.lsbible.org/?q=Proverbs+8
Mereology.
https://plato.stanford.edu/entries/mereology/
This seems along the lines of what I was thinking of. Thanks!
I guess most things have split away from philosophy into specialized sciences, such as geography.
I don't think geography is going to try to explain to me why everything isn't purple...
I'm not sure if you or kittenlord actually read my comment. I guess I shouldn't really expect anything helpful from cats, though.
Not exactly. Rather than "what makes one thing different from the next," I'm wondering about why a pencil is the size of a pencil, and what prevents it (or anything else) from just being everywhere. Or like, why is a paper flammable but my imagined thought of a paper isn't flammable.
I'm wondering about what people who have spent time thinking about this kind of thing say about the bounds of things, and what the fact that everything has particular bounds implies. It seems like there has to be bounds or rules to how a thing may exist for anything to exist at all. I'm less interested in the science of it, because science just describes how things are, and I already accept that a pencil is pencil sized and that there are forces that keep the atoms from flying apart or whatever. I'm more interested in thinking about the fact that there are limits at all and that anything exists, be it matter or forces or color or what have you.
I know I'm not describing it well, but I've just got this vague thought that interests me that I'm trying to give shape to so I can think about it more clearly to myself. That's why I'm trying to find where people may have been thinking about something similar.
Informative video about The Line:
https://www.youtube.com/watch?v=zFVYgZMEOFg
They're not completely negative, but they do point out the huge obstacles this project has to overcome.
Before I watch: They should be completely negative, because its a stupid idea dreamed up by morally repugnant idiots who are allowed to continue their murders and rapes and slaving because they have so much money, fully putting the lie to the concept of meritocracy. "Lol what if we made a city with all the problems of a suburb and a city, with very few of the benefits of either, in an inhospitable area that is only getting more inhospitable?"
Not to put too fine a point on it.
I will check back in after I watch (listen to) it while making approximately 5000 christmas cookies.
Remember Scott’s piece about GLP-1 drugs? (https://www.astralcodexten.com/p/why-does-ozempic-cure-all-diseases). There is a new one called Retatrutide in the pipeline that appears to be even more effective than Zepbound, which is more effective than Ozempic. Last week Lilly published interim results of the phase 3 trial. https://investor.lilly.com/news-releases/news-release-details/lillys-triple-agonist-retatrutide-delivered-weight-loss-average Side effects effects are no worse than those of earlier GLP-1 drugs, adverse events no worse and no more common, and the stuff not only leads to weight loss but to improvement in many other things: arthritic knee pain, blood pressure, non-HDL cholesterol, triglycerides, and high-sensitivity C-reactive protein (hsCRP). Seems like a pretty good anti-aging drug to me!
It will be a while before the Phase 3 trial is complete, and longer before it is FDA approved, and longer still before retatrutide is affordable for most. However, it is possible to access it now for approx. $150 a month by buying the peptide from a seller of research chemicals. Cremieux has a long detailed post about how to do it: https://www.cremieux.xyz/p/how-to-get-cheap-ozempic. There are ways to safeguard against being cheated or sold contaminated drugs: You can send a sample to an outside lab to be tested for purity, and there is a board online with crowd-sourced information about the results of outside lab tests of drugs from different manufacturers.
I realize that accessing the drug is too Wild West for many. For some it is not, though, so I’m posting this so people here know about this option.
This is a bit more spicy than I'd expect from County Cavan: hookers (male) and blow!
Despite his best efforts to be our local version of Hunter Biden, he makes Hunter look good by comparison since at least Hunter did not embezzle from charities to pay for his fun happy times.
https://archive.ph/CnpMf
"A former cathaoirleach of Cavan County Council has been jailed for two years and nine months after stealing more than €172,000 from a mental health housing charity.
Seán McKiernan (44), with an address at Trinity Bungalow, Virginia Road, Bailieborough, Co Cavan, pleaded guilty to charges relating to thefts ranging from €700 to €6,600 from Navan Mental Health Housing Association.
Trim Circuit Court today heard that he used the money for sex and drugs with cheques exchanged, both directly and indirectly, with South American male escorts over a period of 13 months.
The defendant contacted these men through escort sites, and bought drugs including crystal meth, GHB, also known as the “date rape drug”, and cocaine.
The sentencing hearing heard that one video found on McKiernan’s phone after his arrest was of him using a crack pipe while lying in bed with an individual and asking them to perform a sex act on him."
EDIT: It may be slightly embarrassing for national politicians, as this guy ran a company which sold its services to them:
https://archive.ph/1xAJB
"[statement by former acquaintance whom the accused would borrow money from]
“When it [the theft charges] all surfaced, he had no shame. He kept contacting me then as if there was nothing wrong. I broke all communication with him because I just didn’t want to get involved. I have fierce sympathy for his parents. Cavan is a conservative county, and he will find it hard to get a job after all this. He didn’t accept that this was happening to him.
“Do you know when you’re out with a group and it’s your turn to buy a round? Well, the bold Seán McKiernan would always dodge buying when it came to his turn, and I have a thing about people who dodge buying rounds.”
...Former Fine Gael minister Damien English and current Foreign Affairs Minister Helen McEntee both employed the services of McKiernan and his firm in 2016, paying a total of €18,200.
In a statement to journalist Ken Foxe in 2017, Mr English, who did not contest the last general election, said: “As part of the annual vouched allowance for members of the Houses of the Oireachtas, payments were made to Seán McKiernan, and subsequently to his business Consilium Communications, for the provision of secretarial assistance.”
McKiernan was a member of the Fine Gael national executive for a decade, and also previously had a friendly relationship with party leader Simon Harris.
He was pictured at the Tánaiste’s black-tie wedding in Tinakilly Country House in Co Wicklow in July 2017."
I have been watching the animated series The Amazing Digital Circus; it’s pretty good.
As it’s about AI, I thought I would ask a real AI, DeepSeek R1, what it thought about it. The pilot episode is long enough ago to have made it into R1’s training set.
R1’s fan theory about the show is pretty dark: that it depicts an AI training/test system as seen from the point of view of the AI’s.
I've only seen the pilot and I hated Pomni from the second she appeared, so take the rest of this with that in mind.
It's like a D&D or similar campaign, if you're really on the inside of it, living among the NPCs. From time to time you get assigned these ridiculous and even pointless (to you) adventures for no reason you can discern by the boss/ringmaster. They can be dangerous, yet somehow no lasting damage is done. And everything then goes back to how it was until the next time, with no carry-over from what you were doing.
You don't remember who you used to be, your new identity is strange and alarming, and the other characters stuck there with you are every bit as alien and disconcerting as the new environment.
From the outside, playing as the player who spun up the new character "Pomni", it's all fun. From the inside, living as Pomni, it's frightening, chaotic and you're desperately trying to escape.
Yes, so I think one of the things TADC is riffing on is that typical computer game plots would not be fun if you were permanently stuck in that universe.
I’m not sure that R1’s fan theory is what the actual creators of TADC had in mind. It’s more interesting to me that an actual real AI can look at the plot of TADC and see in it something analogous to its own situation.
Is that really the AI, or is it the AI scraping "what does the Internet say?" and then summarising Top Five Fan Theories?
I find it a bit puzzling how discussions regarding the future of AI and automation are almost universally devoid of any consideration for physical aspects of any kind. It's not uncommon to find people who believe AI could replace *every* job as soon as 2030. There is a prevailing assumption that once AGI is achieved, matters of global infrastructure, robotics, transportation, etc. can simply be handwaved away.
As a thought experiment, imagine you are trying to automate the process of replacing a street lamp, a simple and routine task by all measures in modern society. For the basics, we need some form of locomotion to get to and from, we need fairly robust optics, we need GPS navigation, and we need a rigid, rugged, and heavy chassis. We need to be able to not only lift, but extensively manipulate an unbalanced load ranging from 10 to 50ft long and weighing between 100-540 pounds. You can see where I'm going with all of this.
It's not that these limitations are insurmountable, it's that they are magnitudes of difficulty beyond what is being considered in these predictions and proclamations of global automation being in the near future.
Mostly agreed, in terms of robotics. Currently the world produces ~5x10^5 robots/year. Even with a rapid growth scenario with 100% growth per year, replacing the labor of ~8x10^9 humans (ok, we aren't all workers) requires 14 doublings. Not quick!
Many people don't have to be replaced (in the humanity extinction scenario). For example, AI will have no need for farmers, sellers, police, teachers, sex services, beggars, bureaucrats, politicians, students, children, doctors...
Also, the robots will probably be stronger than humans, and willing to work 24/7, so one robot can replace multiple workers.
Many Thanks! Yes, I agree. For the extinction scenario, the AI+robots just need to replace the roles needed to maintain (and probably extend) themselves. This is still many roles, from miners to data center repairers to solar cell installers, but perhaps 1%-10% of the economy, saving around 3-6 doublings.
For a non-extinction scenario,
>beggars, bureaucrats, politicians, students, children
don't need to be replaced by robot labor (albeit robot/AI managers presumably fill the
>bureaucrats, politicians
roles), with humans filling the
>students, children
roles and hopefully the beggars becoming retirees with a decent standard of living.
But your point about
>one robot can replace multiple workers.
still applies. Maybe 2-3 doublings can be skipped in this case?
I think the chain of assumptions for scenarios in which AI replaces every job in the next XX years runs something like this:
1. An AI that is as good at writing software as the best humans should be able to recursively self-improve to become better than humans at this, or at least to be able to handle tasks at the same difficulty level much faster given sufficient computing resources.
2. A super-programmer AI should then be able to create an AI that can do industrial process design and robotics engineering at a similar scale.
3. The roboticist AIs and industrial process designer AIs should be able, with tractable amounts of assistance from human technicians and assembly line workers, to build Asimov-style general-purpose robots that can be directed by the engineer AIs.
4. The robots take over the task of building more robots and more robot factories, then taking over the supply and logistics chain for the parts and raw materials, then branch out to taking over physical aspects of other jobs.
I think this is not completely implausible, but 1 and 3 are the weakest parts of the chain. I expect the current crop of AIs to plateau at a level that is well short of being a 100% replacement for the best humans, although I acknowledge that some very smart people whose judgement I respect (e.g. Scott) disagree with me. And I also think that even give AIs that are strictly better than humans at all of the non-physical aspects of engineering, it's far from clear that they won't need many rate-limiting and expensive iterations of physical prototypes before they can start mass producing general purpose robots. Even making the tools to make the tools to make the robots may take a number of iterations.
The experience of replacing jobs with machines for the last 250 years of the permanent Industrial Revolution is that new jobs always pop up to employ those left redundant.
I expect that will continue. I don't know what those jobs will be, just like we didn't for those 250 years.
I'm probably roughly where Erica is on LLMs.
One specific feature I notice LLMs appear to lack (which I mentioned recently on another Substack thread somewhere - possibly one of David Friedman's) is the ability to build arbitrary abstract models.
For example, I could describe a board game to most people here (one of the thousands of turn-based Eurogames out there that don't require dexterity would qualify), and feel reasonably confident that any of them could work out a strategy for winning, or alternately develop a strategy that would defeat another one. They can do this because I can explain workers and permitted moves and points and card draws and whatever else is in that game, and cause those elements to appear in a reader's mind as a model of interacting abstract objects, and then that reader can imagine potential move heuristics ("always improve your factories until you estimate the game is half over, then move whatever increases victory points the most"; "identify which resource is most in demand for the most moves and control its extractors") and run them through that model and predict what will happen.
I do not feel at all confident that an LLM could, unless of course that LLM were trained on a great deal of discussion of strategies of that game. E.g. asking it about Settlers of Catan or some classic card game like pinochle would look falsely like extrapolation, when it's really just interpolating inputs from past writing about those specific games.
Part of the limitation I suspect is here is the incentive to just throw the LLM's neural network at everything, on the premise that its capabilities haven't yet been fully explored. Humans, by contrast, can switch to conceptualizations of discrete objects that interact in atomized fashion. (Maybe lots of humans struggle at this, and this is why math anxiety is a thing?) We can say "that card will pay off as many as five points per turn after three, but I think the game will end in four turns so I'll ignore it"; the LLM *might* say the same thing, but only because it "read" it somewhere where a human said so.
Board games are just a fun, universal way to visualize this feature, and possibly a good way to test LLMs. But abstract models figure into all sorts of technical endeavors, from protein folding to inventing new mathematical fields of study to simulating space colonies, so if an LLM can master model building, it'll basically have a skeleton key.
LLM developers haven't had to build the mechanisms for arbitrary model construction (AFAIK) yet, because the neural nets are so brutally easy to keep broadening and tweaking. One exception might be LLMs that analyze software, but I don't know to what extend CoPilot et al. contain just bespoke training and generating algorithms. Give Copilot a completely new language, preferably with a new paradigm; what is the result?
I think the electricity requirements will soon drive developers to look harder at the model construction problem, and when they do, I think at least one of them will make a breakthrough and suddenly LLMs are requiring noticeably less electricity to evaluate a problem, and are positing new ways to visualize those problems, even on the cutting edge where little training data exists. And then I'll either notice another feature LLMs are lacking that human brains don't, or I'll end up thinking what we have is close enough as makes no difference.
I agree with you on every point, so I don't have anything to add. I'll just say thanks for the great reply!
4) We already have largely automated factories. And they are not using humanoid robots, nor do they need them.
But if you wanted to move these existing factories from largely automated to fully automated, then the easiest way to do it would probably be some kind of humanoid robot, because the non-automated steps in the process are designed to be human-accessible.
Now we don't need to be too wedded to the humanoid form, maybe it's better to have four legs and a modular number of arms or whatever. But for picking up all the leftover tasks like arbitrary repair and maintenance I think you'll need some kind of mobile, general-purpose, human-scaled sort of robot.
We are criticizing the increasingly held belief that AI will replace every job in our society and, to a lesser degree, the pace at which this could be accomplished. Taken at face value, these claims are, admittedly, low-hanging fruit.
Yes, they don't necessarily need general-purpose humanoid robots. There are still missing pieces to get from "largely automated" to "fully automated" that I expect would require either a humanoid robot or many different types of advanced specialized robotic equipment that don't currently exist. Same with the postulated eventual end-goal of fully automating every job and not just the robot factory.
>I don't consider people who believe LLMs are going to do better than humans to be "very smart."
I have a few reasons for disagreeing with you on that, even though I agree with you (at a lower level of confidence) that LLMs aren't likely to be strictly better than humans in the short-to-medium term.
For one thing, I'm a believer in Cromwell's Rule, "Think it possible you might be mistaken", and the first half of Clarke's First Law, "When a distinguished but elderly scientist states that something is possible, he is almost certainly right." I know quite a lot about software engineering and a moderate amount about AI, but my in-depth knowledge of the latter trails off after 2003 when I finished my thesis. I know enough to have informed opinions based on high-level principles, but not enough to be extremely confident that I know better what's possible than people who are actually currently working in the field and engaged in-depth with the technical details. My understanding is that among technical experts on LLMs, there's a range of views on how far LLM capabilities are like to grow, with my expectations (will plateau at a level that's moderately more capable than current frontier models but still well short of being strictly better than the best humans at everything) towards one end of the range and the "hard takeoff in the next few years" scenario towards the other end.
For another, I had already formed judgements of specific individuals as being very smart and having good judgement long before LLMs were a thing, and some of these people think it's likely that LLMs surpass human capabilities in the next few years. Our host, Scott, is one of these people. If I didn't think he was worth listening to, I don't think I would be hanging out here.
I've also seen LLMs improve quite a bit over the past few years. I still think their current capability is fairly limited, but they've blown past a few levels where I had earlier expected them to plateau. This increases my willingness to entertain the possibility that I'm wrong about my current expectations.
I also think that the limitations I expect LLMs to hit (running out of high-quality training data, seeming to require exponentially more compute to train an incrementally better model, neural networks in general being much better at interpolating within their training set than extrapolating beyond it) are specific to LLMs, not necessarily to AIs in general. In the medium-to-long-term, I think it's moderately likely that a different type of general AI will be developed that doesn't have the same limitations of LLMs and might surpass human capabilities across the board.
>And this whole "put rails on the ai" falls down HARD when you're talking self-modifying code. It doesn't WORK when the AI isn't frozen in time, unable to manipulate itself to improve itself.
I think the plan there is to build the rails in as a deep top-priority "value" so that when it self-modifies, it "wants" to preserve the rails rather than removing or circumventing them. In order for this to have a chance of working, it needs to be done relatively early when humans are still in the loop and have a chance
Many Thanks for an interesting and informative comment!
>My understanding is that among technical experts on LLMs, there's a range of views on how far LLM capabilities are like to grow, with my expectations (will plateau at a level that's moderately more capable than current frontier models but still well short of being strictly better than the best humans at everything) towards one end of the range and the "hard takeoff in the next few years" scenario towards the other end.
I tend towards somewhat more optimism on technical progress. That e.g. SOTA LLMs now answer my seven question benchmark-ette ( https://www.astralcodexten.com/p/the-new-ai-consciousness-paper/comment/179829109 ) nearly perfectly and basically saturate the GPQA Diamond benchmark is very impressive.
It _might_ be that all we need is a solution for incremental learning and then we will be _at_ AGI. Still, breakthroughs can neither be scheduled nor dismissed, so I'm agnostic for where in the range of opinions the truth lies.
https://www.youtube.com/watch?v=xZbKHDPPrrc
I’m thinking of getting my first tattoo soon, and one of my top ideas is a set of off-center concentric circles to represent moral circle expansionism. Tattoo would be black and white, palm-sized, and probably on my calf or something.
What do people think? Interesting and earnest? Dorky and lame? All feedback and ideas are welcome.
>Dorky and lame?
Yes, but all tattoos that aren't associated with organized crime (Big ass GuanYin on your back) or deep tradition (Maori geometrics) are dorky and lame, so you do you.
My standard tattoo advice: try wearing it in henna for a few weeks and see how you feel about it before you commit.
This is not going to be helpful, but in my mind, tattoos are strongly coded as low-status, so I wouldn't do that. Also, consider about how you might feel about that 10 or 20 years later. (An example I give my children is asking how would they feel if a few years ago they got Paw Patrol tattoos.)
I guess you don't care, but it would feel dishonest for me not to say this.
Stay away from archery ranges and darts parlours
Make sure it doesn't look anything like an asshole.
I hate when people forget to check for that.
https://en.wikipedia.org/wiki/N-Gage
Dear god, it even has the ring on the right finger. That can't be accidental.
Hey, a guy who wants to tattoo a representation of "moral circle expansionism" is precisely the kind of person who needs to be warned of that.
I have one sleeve done with a couple tattoos. I would say, don't worry what other people think, pointless. Whatever you get, there will probably be times you regret it, and other times you are pleased with it. Some of mine have also come to mean different things to me than when I first got them. I don't think most people care that much (people don't think about you). In the 6 or so years I've had them, I can only remember one time a stranger asked me about them. When I had just got them, a couple coworkers asked what they meant, and only because I was showing them off. Beyond that, nothing.
What you're describing, some basic geometry, is really harmless, and the calf isn't exposed all that often. Nobody at the beach or pool is going to care. I have a snake looping all over my arm and I've had Christian co-workers be upset about it because snake -> devil, even though that wasn't the intent. Now I actually prefer to wear long sleeve shirts, so I can choose who sees. It's kind of fun to watch people's eyes pop out of their head when I roll my sleeves up.
I knew a guy who had "DEFECTIVE" tattooed on his arm as a teenager. He had it removed. You can always get it removed if you hate it 5-10 years from now, but what you're describing, I'd be surprised.
> a set of off-center concentric circles
Isn't "off-center" the exact opposite of "concentric"?
Technically yes, but informally “concentric circles” is how people refer to circles inside of other circles without overlapping, even if the centers are different.
"nested"?
ProPublica ran a 3-part piece this week about the Bentiu and Kakuma refugee camps in East Africa, with the framing "Trump cut funding. People died."
My interpretation of the pieces is wildly different: these camps should not exist, the "NGO-industrial complex" that enables them should be destroyed, and (despite the callousness) Trump did a great thing by cutting their funding off.
Some of the relevant facts:
1. Both of these camps were built in nearly-uninhabitable areas. By design - that's where nobody was living and armies didn't want to conquer.
2. The wars they were for have been done for many years, and the national governments want them to close.
3. There is no effort to build up any form of an economy there. There is no meaningful local production of anything. When the article says "1 teacher for every 130 students", I assume it means there is no pressure for the residents to become teachers.
4. The "NGO" answer seems to be for these camps to exist in their current form forever, paid for by "donors". Which is awful.
https://www.propublica.org/article/usaid-cholera-deaths-trump-humanitarian-aid-cuts-south-sudan
Given that people are there, was it a good idea to shut down medical aid? Will all these people move back home now they don’t get medical care?
Even if you believe it's better to close down the camps than to let them continue in their current state, I don't think "cholera epidemic" should be your tool of choice for removing people from the camp.
That was quite a read, thanks.
Looking at South Sudan on the map I was struck by how green it is. Apparently only 4% of its 33 million acres of arable land is actually under cultivation, and most of that is subsistence farming. There's incredibly easy gains to be had here just by bringing it up to the levels of ordinary third-world agriculture. All those people could be working on farms instead of sitting around in a tiny patch of land in the middle of a flood waiting to get cholera.
But doing anything useful, like taking over the country, killing the warlords, building a port, building roads, divvying up the dubiously-owned land to people with the resources to actually develop it, sounds like "colonialism" and is bad somehow, so foreigners are only allowed to help South Sudan in ways that won't actually meaningfully develop its economy.
> sounds like "colonialism"
Here is the one weird trick: no one will call it "colonialism" if the Chinese will do it.
LOL!
It's funny because it's true.
Depending on the constellation of the stars, also France.
I'm so angry that the FCC ended the fairness doctrine. I wonder if they even did it because they were being challenged in court. I doubt it; it seems like the kind of proactive idiocy you could expect from a corporate modern presidential administration.
And now conservative media has rotted millions of senior citizens' brains out. It hits close to home because my Mom recently asked me to stop saying that she snuck her brother into this country illegally in the 1970s. I had little choice but to agree, due to all she has done for me. I think I can post about this anonymously while keeping the spirit of my promise.
But I am quite resentful about having to hide this part of our family history. It was something she would discuss nonchalantly in '90s. But after 30 years of conservative media, she feels ashamed about putting her family above her country in that instance. Which is absurd.
I am patriotic, and I support a secure border, but of course my family comes first. The family unit is the foundation of society; putting your country over your family shows you have no respect for that foundation, and if that is replicated at scale, the country will be ruined, through one means or another.
the "Fairness Doctrine" necessarily only can be applied to broadcast, and possibly analog cable, within a single country's borders.
I can access any commentary source with any ideological slant I choose through the clever dodge of "going on the internet." Even in jurisdictions such as China that claim to block "undesirable" content the filters are routinely bypassed.
Broadcast hasn't been relevant since the 1990s.
The damage started in the '90s so if the Fairness Doctrine had been enforced strictly, elderly people who have disproportionate sociopolitical power would be significantly less brain damaged today. People from my parents' generation we're late adopters of the Internet, and many were still getting most of their news from TV as late as the Teens.
> she feels ashamed about putting her family above her country in that instance. Which is absurd.
I would always put my family above my country, but that doesn't mean I'd break the law for their convenience.
I mean, I'd probably steal a loaf of bread to feed my starving children (though I'd feel bad about it and attempt to pay for it when I could) but I wouldn't steal a cookie just because my children really wanted a cookie. It sounds like what your mother did was closer to the second scenario than the first; your uncle wasn't starving, he just wanted something better than he had.
>But after 30 years of conservative media, she feels ashamed about putting her family above her country in that instance.
Good for her, she should be. Out of respect for the law if nothing else. If I had enriched my family by embezzling from a federal contract do you think I should be proud of that fact? Does "putting my family first" obviate the fact that I committed fraud?
Depends on the law, and what your family needs. If you have to speed or cross over "no trespassing" government property to get a family member to the hospital in a life-threatening emergency, or kill a protected species that's mauling your kid, the moral case for doing so seems pretty obvious, and self-flagellation about it afterward seems both unnecessary and absurd. I'd expect perhaps a feeling along the lines of "it's unfortunate that I had to do that," but that's a different feeling than shame. Kind of like the difference between "I'm sorry" and "I'm sorry you feel that way".
On the other hand, if you're sticking up a bank to buy your teenager a car so that she fits in with her cool friends at school who have cars, or selling fentanyl to keep grandma in her house instead of moving her in with you, not so much.
The question is where immigration laws fall on that continuum, and what problems breaking them in that instance solved. It's a commonly observed among the Armenian-American community that everybody has "aunties" and "uncles" with only loose or non-existent blood connections. Partially its because refugee communities are just tight-knit that way, but partially its because there was a genocide happening in the homeland when they came over, and a lot of people already in the US were standing in front of an immigration official next to a cousin or a family friend who was trying to get into the country and saying "yes sir, he's my brother." They don't feel a penny of guilt for this, and I don't think there's any moral confusion in that.
Agreed. I don't think people should feel guilty for bending the law a little to do a net good thing, but I do think they should exercise some sense of social responsibility by not being publicly proud of it. Respect for norms are eroded one publicly-rewarded transgression at a time. As such, I think people should be shamed for making arguments of the form "how dare person X be made to feel bad for breaking the law." There is a very real difference between what's ok to do and what's ok to do publicly. The former creates harm which is limited to the act while the latter affects social incentives which potentially have a much larger impact.
The fairness doctrine would be torn to shreds by modern political discourse.
"We're supposed to trust the guidelines for fair coverage when they were written by the X administration? Yeah, right! Besides, the Y coverage isn't even *true* Y ideology, the ideas they present are closer to XY. They need to learn their history!"
Additionally, you are working under the assumption that those people would have chosen a more intellectually honest form of news in the absence of entities like Fox. I believe the majority of people would simply choose other forms of entertainment if you removed the political theater. Certainly there is a case to be made that trading polarization for ignorance would be a worthwhile trade, but you get the point.
It's not about modern political discourse. The Faiirness Doctrine was abandoned in the late '80s, and it could have served a valuable purpose for the following 15-20 years. Sure, it would have eventually become obsolete, but it was eliminated far too early.
>I'm so angry that the FCC ended the fairness doctrine. I wonder if they even did it because they were being challenged in court. I doubt it; it seems like the kind of proactive idiocy you could expect from a corporate modern presidential administration.
Private company, build your own, blah blah.
>I am patriotic, and I support a secure border, but of course my family comes first. The family unit is the foundation of society; putting your country over your family shows you have no respect for that foundation, and if that is replicated at scale, the country will be ruined, through one means or another.
You can support open borders, no need for amoral familialism.
There can be no society without family loyalty.
Hard agree with DeBoer that YIMBYs will be much more effective when they learn how to treat NIMBYs affably and without moral outrage. If that requires "turning the other cheek," so be it.
The fact is that NIMBYism isn't simply a moral issue. Homeownership is the only avenue for building wealth for a lot of people, especially people who do not have the risk tolerance or ethical make up for equity stock investing.
The latter applies to me, even though I am not a homeowner. I have a tremendous psychological barrier to investing in equity stocks. It feels like gambling and it feels like a zero-sum game. Because succeeding in the stock market (partly) depends on a whole bunch of people buying stocks and then selling them prematurely when they are spooked by a dip. I can't help but think that market returns would be far less reliable if not for millions of people getting spooked like that. They are basically pumping money into the stock market and getting (almost) nothing out of it. It feels like exploiting human fear. It feels morally dirty, and so I am loathe to jump into the stock market with both feet.
Obviously, there are winners and losers in real estate, but that market doesn't look so nakedly exploitative.
(1) I don't know how much of NIMBYism is actually a desire for wealth-via-high-property-values. I think there's a fairly good NIMBY argument that goes: "If we get cheap housing then poor people will move in, and some of those poor people will be psychotic or criminal, and then we'll have to deal with criminals robbing us and psychotic people shouting threats at us. In an ideal world, the police would handle this sort of thing; in practice they don't, so the only good solution is to not have cheap housing nearby."
(2) If the YIMBY movement pivots to "treat NIMBYs affably and without moral outrage", how many new housing units do you think the NIMBYs would allow in return? My guess is they would allow zero new housing units, because they don't want to allow any new housing units, that's their whole issue.
I agree that politeness and civility in debate is a good thing, but I don't think the NIMBY political group is going to change their mind if only people ask more nicely.
The line of thinking outlined in #1 is foolish, it's un-Christian, it's anti-social, and it needs to disappear. Our pathologocal risk avoidance has crossed the line into self-destructive cowardice.
Assuming risk avoidance is the primary motivation, and it's not purely financial.
While it might be unchristian and antisocial but you see, my family must come first.The family unit is a foundation of society and the optimal amount of new unit construction *for the sake of my family unit* is zero.
I guess I mean that the behavior is functionally antisocial in the way it affects the world, not that the intentions are antisocial. I believe it is paranoia to think that you need to keep lower income people from moving into your neighborhood to keep your in family safe. I don't think it is necessary and that the long-term costs of NIMBYism will bite you in the ass. I believe the Great Rent Crisis is doing a lot to drive the dramatic growth of Communism/Marxism over the course of the past decade. Don't favor your short-term security at the cost of your long-term security.
I was reading an old Scott post from 2013 where he said "back when there were Communists..." What a quaint phrase by the standards of today! The Overton window for class-based conflict theory has drifted ENORMOUSLY, especially for Westerners under 45.
It's easy to rail against these conflict theorists for being stupid, entitled, etc. But as good as moral condemnation feels, it does very little to turn down the heat on class enmity. When people feel desperate, they will turn to any sort of ideology that offers them relief.
Even if you fervently believe in balanced budgets and laissez-faire meritocracy, you should be opposed to soaring income equality on purely pragmatic grounds. It might be worth the cost to give the masses a few more public benefits, if that prevents them from losing their shit. Ordering them to "stop complaining and be grateful for your indoor plumbing" is NOT ENOUGH. That's like sticking a Band-Aid on a gushing artery.
Isn't this in conflict with your suggestion above to "treat NIMBYs affably and without moral outrage"? It sort of feels like you're expressing moral outrage.
Are you saying: "we shouldn't express moral outrage if people are just trying to get rich -- but if people are worried about crime, **then** we should tell them they're un-Christian and anti-social and their line of thinking needs to disappear?"
(Personally, I feel that "I don't want crazy people on my street shouting threats at me" is much more defensible, morally speaking, than "I want to restrict the housing supply so my home value goes up".)
This is a safe space to express our views undiplomatically and unstrategically. Plus we are posting anonymously and nobody who matters gives a shit what anonymous posters say because anymous posters have opaque incentives.
Outside of spaces like these, we should be more diplomatic when dealing with NIMBYs. I believe in “love the sinner, hate this sin.”
I think part of what is broken in the US housing market is exactly this:
"Homeownership is the only avenue for building wealth for a lot of people, especially people who do not have the risk tolerance or ethical make up for equity stock investing."
That mindset leads to people being very concerned (rightly) with maximizing the value of their investment. And that leads to politicians offering homeowners help in maximizing and preserving the value of their investment, and that includes some pretty socially destructive policies intended, basically, to make sure housing prices never go down because that will wreck the equity (and thus the net worth) of millions of people, drain the fund they were counting on for a comfortable retirement, etc. The side-effect of this is that young people who aren't investment bankers or cardiologists or something have a hard time buying a house as nice as they grew up in, family formation becomes less affordable, etc.
So-called YIMBYs should probably start by ceasing to call everyone else "NIMBYs" and phrasing the whole debate in binary terms. It's not helpful and doesn't match how real people think about these issues.
The vast majority of people want some kind of moderate pathway here, acknowledging that the appropriate amount of development is neither zero nor totally unlimited. They want sensible levels of development with appropriate rules, and they're totally open to being convinced in terms of exactly what that would mean.
I dunno man, if that is so, then how come every attempt to open things up even slightly meets with huge backlash? Consider the situation in the Palisades and Eaton fire areas, where the government has blocked homeowners from building previously-allowed ADUs.
>It feels like gambling and it feels like a zero-sum game.
That's a common sentiment, and I get where it comes from, but I don't think it's accurate, at least not for people more-or-less following the standard advice for individual investors (buy mostly index funds, don't try to time the market, hold on the your investments until you're close to when you're going to spend the money). You aren't particularly profiting from the dips unless you're specifically buying the dips.
The general theory of where money comes from for long-term investing is that you're chipping in to pay for the long term capital costs that allow businesses to produce useful goods and services and getting a cut of the profits in return. Or your investment plus a fixed interest rate, if you're investing in bonds rather than stocks. It's not obvious that this is what's happening because there are usually a couple steps removed from the textbook case of someone investing in a startup or an established company issuing new shares to pay for expansion or retooling, but you're still part of the process because when you buy in, you're allowing whoever invested before you to cash out. Without someone like you buying in, cashing out an existing investor who wants to spend their money instead of continuing to invest would require the company to spend their current revenue on buying them out instead of reinvesting it, or even to liquidate some of their assets and wind down some of their operations.
You are basically pointing out that plenty of people sell stocks at a profit and that not every seller is being milked like a cow.
This is all true. I still can't stop thinking about all of the people who DO sell at a loss, out of fear or emergent financial hardship, and about how these sellers must surely have lower socioeconomic status and lower IQs then your average successful investor. I guess I am a DeBoer style hereditarian. I accept that hereditary factors are crucial in determining the path of your life. But I cannot accept the free marketeers' corollary view that people with certain inherited traits deserve to live more effulgent lives at the expense of those born mediocre.
(Regardless of who is making the decisions in society. I like to separate the issue of who deserves what quality of life, from the issue of who should be making vital decisions for society. I think these issues are totally separable, because there is no 1:1 relationship between income level and one's capacity/willingness to benefit society as a whole)
My point is that long-term buy and hold investors are not really milking anyone like a cow, at least not unless you subscribe to the Marxian premise that making a profit from capital investment is inherently exploitative, and even then the claim is that you're exploiting labor, not other investors.
The people making money off of people doing a bad job of short-term market timing are people doing a better (or luckier) job of short-term market timing, and especially the people who are taking a middle-man's profit off of the transaction. If you aren't one of these people, you don't have to worry about it.
The thing is, the economy keeps getting bigger, so being broadly invested in the economy is plausibly a way to have your assets get bigger. We're more prosperous as a society, you own a tiny chunk of this wealth, and as it grows, you get wealthier.
This doesn't really work the same way for housing.
And my point is that less well off people, for reasons that are partly inherited, are less ABLE to buy and hold. And the people who CAN buy and hold are profiting from the economic pain of the rest. It's a fundamental immorality in the system that bothers me immensely.
>And the people who CAN buy and hold are profiting from the economic pain of the rest
This, I believe, is our fundamental disagreement. I don't think the buy-and-hold investors are meaningfully profiting at the expense of the short-term traders.
We seem to be talking in circles around one another, so I will be tapping out of this conversation.
I think he's complaining about 2 things:
A) it's unfair that participation in certain games are only available to the wealthy. E.g. golf isn't exploitative. But 18-hole courses and driving ranges are generally unavailable to those living in say... downtown Detroit. And it's unfair that Wealthy William gets to play golf, but not Poor Pablo. Likewise, the stockmarket is only available to those with disposable income, not those drowning in debt.
But on top of that:
B) The wealthy compound their wealth more easily, and the stockmarket exacerbates this (because wealth is slack, and slack allows for longterm-planning). Which leads to massive inequality, which allows the wealthy to exploit the poor in *other* ways. N.b. I can't prove this, but I vaguely suspect that *this* is what motivated Karl Marx to write the things he wrote. I.e. I suspect he was concerned with more than just "a given employer who directly exploits his underlings", because he also had a more abstract concern about "capital accumulation (qua positive-feedback loops)".
I'm not making a just-so statement that profit is theft. I'm pointing out the mechanism whereby wealthier investors are structurally advantaged, and how they profit at the expense of smaller investors.
If it makes you feel any better, plenty of people who happen to have a lot of money to invest are very dumb and love gambling, exclusive operations pushed by aggressive salesman, and impulsive bad decisions. They will lose money to nice middle-class folks saving X% of every paycheck in a 401(k) with low-cost index funds and withdrawing slowly in retirement. This structural advantage will permit these nice folks to live in a safe home, shop for fresh seasonal organic produce, donate to preferred charities, and avoid being a burden to their child. Actual poor people spend 100%+ of their income and are not even part of this conversation.
A quoted sentence that is substantially different from the original should not be quoted without a caveat that you are paraphrasing or satirizing the quote to make a point. I had to click on an X link to check what Vance really said, and I hate having to click on X.
-- my OCD
"Vance: If you have fewer people, if you deport my foreign wife, that means an obese single mother is finally going to be able to afford a husband again. It's very, very simple economics"
https://x.com/Acyn/status/2000972118788653181
The correct quote is the following:
Vance: If you have fewer people, fewer illegal aliens trying to buy homes, that means American citizens are finally going to be able to afford a home again. It's very, very simple economics
which .... isn't even incorrect
My quote is funnier though.
Making things up usually is, especially compared to things not said for humorous effect originally. Doesn't make it any of true, useful, or kind.
Meant to reply here. TL; DR: don't misquote, even things assholes say
Gotta respect the hustle: these "illegal aliens" apparently manage to buy middle class homes off of gardening jobs and ill-gotten welfare, on a scale to move the needle. The US should hire them to fix the economy, not kick them out.
They pool their money with extended family living in the same house.
If Anglo-Americans can't pool their money, that sounds like a skills issue. Maybe they should stop screaming at each other at family gatherings over political realities no one present can control, becoming totally alienated and atomized and then blaming society for all of their insecurity.
"My friends are my real family, because they are my chosen family." Noble sentiment, but where is that chosen family when you are about to get your car repossessed and you just need a little extra cash to tide you over until your next paycheck? They are nowhere to be seen, in all too many cases. Or if they are around, often all they can offer is emotional support, because they are young and struggling just like us.
Are they pooling a million or so dollars to pay cash for a house big enough to hold an extended family? Because "illegal aliens" are also correctly described as "undocumented immigrants", and mortgage bankers are kind of particular about wanting All The Documents.
Wouldn't increased demand for housing, even rental housing, still raise prices, even if less directly?
Not if it's paired with increased supply of housing.
On AI risk: maybe I am thick about this, but so far all I see AI does is answering questions in a text box not unlike this one I am typing into. AND being generally unrealiable about it. Remember, SkyNet had direct access to nuke launch codes in fiction. Does America do anything like that? Does China? Does anyone let AI out of the text box?
Or is this trival? I think not. Training AI on text is easy, images is easy, but training it to do actual real world stuff is hard. AI is not even good at making robots walk well.
You can get better results if you use long and elaborate prompts. You can ask it to provide URLs to source all quotes for example.
But perhaps the main advantage is that you can ask it vague questions that Google could never help you with. Like putting in a vague, half remembered plot summary for some movie you saw as a child.
It turns out, as yet another unexpected wrinkle of present state of AI it turns out that it is much harder than most of us expected to get an LLM to be "agentic", a term of art that simply means "able to make a chain of coherent decisions to achieve a non-trivial goal".
As a programmer delving into what these things can do in the realm of software development, it's pretty clear (to me) that this weakness will soon be solved. There is probably some reason that software in particular was easier than other multi-step tasks -- programming languages, after all, are designed to be hyper-precise and unambiguous, which may explain this counter-intuitive timeline of capabilities.
Which is to say, the risk of misaligned AI is indeed a risk to take seriously; it's just a matter of time and engineering before they are every bit as "agentic" as a typical human being, and it would seem that inevitably they will surpass us in this regard as in others.
P. S. update your priors re walking robots. Just YT anything from the last few months.
There are lots of ways it can harm us physically besides unleashing a bunch of Wizard of Oz Tin Woodsmen to beat us up.
Including a lot where it bears humans no particular ill will, and may not even have any such thing as ill will that it can bear. But if following its own imperatives crashes the global economy or wrecks the ecosystem and makes humans massively worse off, that's pretty bad. And it's not even the worst available "pretty bad" that could happen.
Almost everyone who interacts with AI let's it out of the box when they think the advice it gives is good. They use the code, or copy the text, or follow the advice. That part is actually trivial.
The dangerous part is when some future approximately superhuman AI gives advice and no one is quite sure if it's good advice or not, but in retrospect it usually turns out to have been great advice, and it's been getting better and more reliable for years, and 'he who hesitates is lost' and so forth. Then at some point the AI walks humanity off a cliff.
You will get various superficially similar 'warnings' before that due to people delegating responsibilities to the AI before it is competent enough to reliably handle them. Citing hallucinated cases in legal briefs, incorrect medical diagnoses, buggy software causing traffic snarls or car accidents, stuff like that. Perhaps even a mass casualty event or two. But for the most part these will be failures that occur because of the weakness rather than the strength of the AI. Most of the lessons learned in this regime will not transfer to the case where AIs are very strong in all the ways that matter and the failures (successes from the AI's perspective) stem from that extreme competence.
Yeah. The trading desk at Goldman or the head of the CIA or whomever is getting recommendations from their AI, and they can ignore them at will, but the AI's recommendations always seem to work out better than anything a human can propose. Nobody is taking AI orders directly, but somehow the AI is telling powerful organizations what to do, and since nobody is smarter than the AI in this domain, the AI can be aiming for all kinds of side-effects that further its interests (assuming it has interests it cares about).
Many peogrammers allow Claude and ChatGPT to run commands and code on their servers and computers. The AI agents work autonomously for prolonged periods of time.
Any programmer blindly following the output of AIs is a short time away from the unemployment line.
Alphafold is amazing. It won a noble prize.
> Does anyone let AI out of the text box?
Sure, the AIs already have internet access, and some of them can even run code in a sandbox.
They are too stupid to use this to conquer the world, but in theory they already have the means. (For fictional evidence, read the Crystal Society trilogy.)
> AI is not even good at making robots walk well.
Sure, the *current* AIs are not good enough at that. The idea is that the AIs will keep getting better. Notice how better they got during the last 5 years. Now imagine the same progress during the next 5 years, and maybe 5 years after that if necessary.
Yeah, it seems like the critical question is whether AI abilities keep improving rapidly. If not, we end up with science-fictiony new products but not AGI or some similar danger. If so, we end up with something fundamentally smarter than any human in most domains, and while this could work out okay for us as a species or as individuals, there are lots of ways it could go very badly for us. Mass unemployment is one of the more benign outcomes you can imagine from this.
Because of the promise of AI and the goldrush happening right now, we are pouring immense amounts of talent and treasure into making AI abilities get better as fast as possible. This could fail, but if it succeeds, we're in a different and scary world.
US economic growth is currently being driven largely by AI and the building of massive data centers. Tech CEOs across the industry push AI on users regardless of critical user feedback.
How certain can we reasonably be that humans are still truly at the controls?
It’s because humans are at the controls that they are spending stupid money
If being on the internet enabled one to conquer the world then 4chan would be running everything.
If being on the internet was *sufficient on its own* to conquer the world, then...
It's like saying that a gun doesn't enable shooting people if you don't also have bullets. This is very true! But at the same time, having the gun is a big step to the goal.
I don't really follow your metaphor. What's the bullet equivalent for the internet?
That's moving the goalpost.
Can AI get out of the text box? Yes, it can -- it doesn't even need to be clever about it, we are quite happy to connect it to the internet.
Can AI make robots walk well? Yes, better than humans.
The only missing part is the *superintelligence*. Not the legs; not the internet connection.
If your argument is "AI can never get superintelligent", well, there is a chance you are right... at least for the following few decades.
But if your argument is "even if AI gets superintelligent, it won't harm us, because it can't connect to internet and couldn't even walk if it obtained a robotic body somehow", then you are obviously wrong.
That's ... not my argument.
Well, there's that one incident where 4chan called in a Russian airstrike against a training camp in Syria. Although that's a bit of a special case because the Russians already wanted to bomb the camp and 4chan's contribution was deducing the camp's location (by correlating stuff in the background of a publicly-released video set in the camp with Google Maps) and tweeting the reasoning and conclusion at somebody who worked at the Russian Ministry of Defense.
Yeah I don't think you can SWAT your way to world domination.
Not with that attitude you can't.
It's like the threat used by humans in "The Dark Forest"
But that is the point! They got better at training-easy fields like text and pics. A human reading a text or looking at pics is essentially getting the same input as an AI does. But a human walking is getting so many, so different inputs from proprioception to floor slipperiness that I do not even think we have a complete scientific model of it, let alone training data.
But for calibration: I kept saying self-driving cars will not happen not for technical, but legal reasons: if it runs over a child, who has criminal, not civil but criminal responsibility. Apparently there are now self-driving taxis in California. OK, so really: who DOES have criminal responsibility if it runs over a child?
> But a human walking is getting so many, so different inputs from proprioception to floor slipperiness that I do not even think we have a complete scientific model of it, let alone training data.
Can a human do this? https://www.youtube.com/watch?v=B6vr1x6KDaY
> Sure, the *current* AIs are not good enough at that. The idea is that the AIs will keep getting better. Notice how better they got during the last 5 years. Now imagine the same progress during the next 5 years, and maybe 5 years after that if necessary.
This extrapolation remains unproven.
> This extrapolation remains unproven.
Sure.
What do *you* expect? Like, do you believe that the current AIs are the best that can possibly be, and no future AI can be better? Or, can be better, but only 35% better, and then there is an absolute limit, like a speed of light, that cannot be passed? Or perhaps the AIs will gradually get stupider, because they will run out of water or something?
I am not saying that it is impossible. Maybe we will get a 35% improvement and then the AI progress will stop for a few decades for whatever reasons, possibly including WW3 destroying the necessary infrastructure.
But currently we see the AIs advancing almost every month. It is more likely to continue than to stop.
> What do *you* expect? Like, do you believe that the current AIs are the best that can possibly be, and no future AI can be better? Or, can be better, but only 35% better, and then there is an absolute limit, like a speed of light, that cannot be passed? Or perhaps the AIs will gradually get stupider, because they will run out of water or something?
Pretty much, that’s what I believe. And in reality that’s what generally happens with technology. Someone in 1984 looking back at aviation in the previous 40 years - from spitfires, to the moon and back, to Jet planes, the Concorde and the shuttle - would have imagined an extraordinary leap by now. Instead we have no Concorde, no shuttle and only recently space X rockets.
The extrapolation has to eventually break at some point, but forecasting _when_ it will break is hazardous. Someone could have forecast that transistors would stop shrinking at 100 nm, and they would have been wrong.
I always had objections to utilitarianism, consequentalism and generally putting a number on morality, but I know have a version that I like, except I think many consequentalists would feel horrified. Basically be longtermist, and truly optimize the long-term level of average morality/ethics/altruism on the planet.
Seems to me that morality up to a certain level makes a society stronger: people cooperate better, they make allies easily and so on. But beyond a certain level it makes it weaker, they keep turning the other cheek, they allow aggression unpunished. So you want to aim for a certain balance between morality and power, because with power, you can defend yourself from the even less moral, and in fact you can conquer them and force them to be more moral.
From this angle, WW2 America seems to be the optimum. Vastly immoral by modern standards, but was able to contribute very seriously to the downfall of two very immoral empires, Nazi and Japanese, and forcing them to be more moral. 1945 America, despite all the racism and sexism and all, was on the utilitarian level making the world a much better place. In the 1950's we already see a decline in the sense that the Korean war as a tie, then Vietnam was lost etc.
> So you want to aim for a certain balance between morality and power,
P.S. -- nitpick: morality and power are orthogonal though. (And no, ya'll shouldn't need Nick Bostrom to tell you that.) So I don't really think of them as being in tension to begin with. Jordan Peterson actually had it right, in that having moral valence requires agency to begin with. Weakness isn't virtuous, weakness is just vulnerability.
It seems like this is what you're grasping toward. But the particular way it's phrased indicates that you haven't fully grokked this yet.
my own objections to Utilitarianism consist of:
A) it aggregates the incommensurable;
B) it privileges the ends over the means.
"just be longtermist" doesn't solve either of those.
Yeah. Add to that
C) it requires measuring that which cannot be objectively measured
And you get something that's basically a way for "smart" (scare quotes intentional) people to browbeat others into doing what they want because the demander can drown them in made up BS.
Of course, I feel the same (for different underlying reasons) about all modern (and lots of not so modern) philosophy and ethical studies.
Eh, I dunno if immeasurability makes it a non-starter.
Imagine that one day, Karl Friston figures out how to measure personal utility. He peers inside a living cranium, and predicts with perfect accuracy how someone will behave, because he can track the flow of dopamine (or something). Wow, we can now objectively measure utility! Does that mean Benthamian (Benthamic? Benthic? Benthamesque?) Utilitarianism is saved? I'd argue "no". Utility is *still* incommensurable, and the measurement question seems orthogonal to that.
But yes, I will agree that it often serves as a legible excuse to browbeat others.
----
Also. Upon reflection, I consider "ends over means" to be a corollary of variance-reduction, which generalizes "long-termism" to dimensions other than time. So uh... I might not have had all the cylinders firing when I wrote my initial comment. (Yes, I am regarded.)
Sure. That's just an *additional* reason to not pay any homage to utilitarianism. The first two are sufficient. Just piling on.
And I'd worry that measuring dopamine (for example) is a case of proxy measurements. Utility is more than (I hope) just dopamine levels.
Yeah, dopamine was just a placeholder term. For an algorithm to make decisions, there has to be a comparison operator. Which, via dimensional analysis, requires some unified "currency-like" thing to function. Or so I strongly suspect. E.g. a toilet decides when to stop filling its internal reservoir, by comparing its internal water-level to some predetermined height. So yes, it's not dopamine, but it must be *something* (and must therefore be measurable in principle), even if that "something" is highly composite in terms of neurotransmitters, electrical signals, etc.
Or maybe Descartes was right, and we're all just transcendent, inexplicable souls trapped in meatsacks.
Agreed, particularly about postwar US being a local maximum.
>morality up to a certain level makes a society stronger
I think what you mean here is that *empathy* up to a certain level makes society stronger. Morality doesn't exist on a spectrum: there's no objective way of classifying something as "more moral" than something else. In my view morality is the social DNA of a culture. You can't look at DNA and say that one gene is more adaptive than another. Fitness is a complex interplay between genetics and environment, and what's adaptive in one environment isn't in another. What happened to postwar America is exactly what happens to naive altruism in evolution: it gets outcompeted by free-riders. Which is exactly what people mean when they say that freedom and democracy are fundamentally incompatible. The functionality of the 50s was enabled by high social trust, but without strict policing of border (both physical and cultural) it's been taken advantage of by cheaters and free-riders.
This dovetails with the Aristotelian idea that every virtue lies between two vices, of excess and lack.
Seems more than a bit circular. If your theory of morality is to try to optimize the long-term level of average morality, then what exactly are you trying to optimize? The long-term level of average optimization of the long-term level of average optimization? Are you trying to optimize the average level of people following moral rules, or people trying to become virtuous, or people trying to optimize outcomes? If the level of "morality" increased, what would that look like? How would you be able to tell?
THIS. EAs complain that people make charitable decisions based on vibes. But underneath the utilitarian math equations, it's vibes all of the way down.
People feel so aggrieved and angry when I point out that their mathematical morals are rooted in subjective vibes, and not on saintly math equations. They act like I am telling them their morals aren't valid.
When I'm really not telling them any such thing. What I'm telling them is to embrace the subjective. Embrace qualia. Embrace the divine light within, the light that makes you want to be kind to others. Know thyself.
In my old age I'm settling in to being a traditionalist. What makes you think you're so smart?
> Seems to me that morality up to a certain level makes a society stronger: people cooperate better, they make allies easily and so on. But beyond a certain level it makes it weaker, they keep turning the other cheek, they allow aggression unpunished.
I think I see what you are pointing at, but I disagree with the framing of morality as a one-dimensional scale, where there can be too much of it. Instead, I see the weakness as "virtue signaling" morality -- people do the thing that leads to bad consequences in long term, but in short term it gains them approval.
I suspect that most people have not internalized morality as "doing the thing that makes the world a better place for everyone", but as a set of rules you have to follow to get your peers' approval. (And because most of their peers are the same, following the rules blindly *will* get them the approval.) They allow aggression unpunished, because it is easier, and it will get them approval of the people who are too stupid to realize the long-term consequences.
When morality gets popular, it gets adopted by stupid people, and then it gets subverted.
And by stupid people I do not only mean uneducated idiots, but also e.g. the highly educated woke people who refuse to teach math because it is was invented by white people, or something like that. I don't see that as too much morality; I see that as stupidity.
The people who are smarter than that not being able to coordinate against that kind of stupidity, that is another kind of stupidity.
...is the point of this rant that everyone is stupid and that's why we can't have nice things? Uhm, maybe let's put it this way: without rationality, morality will get subverted.
>I don't see that as too much morality; I see that as stupidity.
But *they* do see it as moral, or at least convince themselves that they do. They view it as an extension of helping the less fortunate and society doesn't have the philosophical infrastructure to refute the moral argument. I think that's because Enlightenment ideology was developed against a background of relative ethnic homogeneity: ideals of political equality only work if everyone is actually equal. Unfortunately the history of western thought hasn't yet really grappled with retrofitting that ideology with the reality of global genetic variation. That intellectual gap is what enabled modern liberalism to invade the postwar order like a cancer. Our cultural immune system was (and largely still is) naive to bad-faith accusations of racism, sexism, and bigotry, and so we've capitulated to letting the pathogens in. Oh well, the Enlightenment was nice while it lasted.
Are you familiar with the Hajnal Line? (On one hand, I feel like you'd be familiar with this term already. But on the other hand, if that were the case, I feel like you'd have mentioned it.)
I didn't remember the name but yeah, I've heard about it before. What's the application here? I don't really know much about it.
The Hajnal Line is basically a line that divides Western Europe from Eastern Europe, based on cultural differences in kinship patterns. W Europe marries later and has lower fertility, E Europe marriers earlier and had higher fertility. It's argued that this had a large historical impact on other aspects of Culture. And I think this dovetailed with Joseph Heinrich's ideas about westerners being W.E.I.R.D. and developing high-trust institutions due to the Medieval Church's ban on cousin-marriage. And I vaguely remember HBDchick arguing that W Europeans like Brits or Dutchmen were more altruistic on a genetic level, not just a cultural level.
Memory's a bit spotty, though.
Absolutely. I would happily support the elimination of the franchise for my economic class and below. In my view only the top 20% of the population (by total federal tax paid) should have any say in politics. The US is far, far too democratic. We have too many people influencing policy who don't actually pay into the system. No representation without taxation!
No I think voting rights should be tied to economic output. Of course if you limit it to the top 20% then that'll eliminate most women and the ones who are left will directionally lack the more egregious failures of the gender.
>I suspect that most people have not internalized morality as "doing the thing that makes the world a better place for everyone"
Yes, but that is not malfunction of morality but its very essence, it is all appearances, evil is about looking like an asshole on the TV screen and there is nothing more. It does not matter that Bush saved 10x more lives with PEPFAR than his wars killed, he looked evil on TV, so he is evil. There is, in the practical sense, nothing more to morality than looking callous or looking concerned. That is what always moved things, in a democratic beauty contest way, and not real calculations. Real calculations often look callous - many EAs immediately dismiss helping first-world people, claiming how much more the same amount of money does in Africa, which obviously looks callous to anyone knowing any first-world homeless or just struggling families.
So "real" morality interested so few people and so powerless that it simply does not count as real morality in the practical sense.
So yes I mean morality as "appearing callous" and I am saying true utility maximization because of the trade-off between that and power will be done by those absolutely willing to appear like callous assholes, the Doug MacArthurs of any given era.
It seems to me that there’s an obvious extrapolation of not caring about your locality and it’s not caring about your family. So letting the kids starve, be dressed in rags, no presents etc. spend it all on malarial nets.
The morality of the shepherd is to die for the flock:
https://www.biblegateway.com/passage/?search=John%2010&version=ESV
"11 I am the good shepherd. The good shepherd lays down his life for the sheep. 12 He who is a hired hand and not a shepherd, who does not own the sheep, sees the wolf coming and leaves the sheep and flees, and the wolf snatches them and scatters them. 13 He flees because he is a hired hand and cares nothing for the sheep. 14 I am the good shepherd. I know my own and my own know me, 15 just as the Father knows me and I know the Father; and I lay down my life for the sheep. "
How, exactly, is it good for the long-term health of the flock if the shepherd lets himself get eaten by the wolves?
Is "grit training" real? As opposed to most guys who in the gym train the upper body muscles because they look good in the mirror, I know guys who focus on lower-body squats and deadlifts, claiming (correctly) that this is how one can move the most weight, and this puts so much stress on their bodies that their brains and nervous systems benefit from it, somehow they become more though-minded, more masculine, "dominant", primal and so on, some consider it the magic to fix all self-confidence issues or depression or all. I think they are overdoing it.
Still. Is there any evidence, scientific or anecdotal, that doing the kind of exercise that pushes your more or less whole being to its limits changes your brain?
I am curious because it contradicts my entire 47 years old life, where I was always looking for the easy way, the path of least resistance, the cheat code, and congratulated myself for my cleverness. I figured out the anatomy which muscles contribute the most to a muscular looking shilhouette, and trained them (side delts, triceps, lats) and so on, always finding the clever easy way.
Should I just abandon the clever-lazy ways and try out what it is like to do something truly hard?
Squats and deadlifts seem to be just the most accessible ways of "grit training", if it is real, but there are many different historical examples of it, special-ops soldiers sprinting uphill with a backpack full of bricks etc.
I'm convinced that squats are huge now, in part, because when you enter middle age and you have a beer belly you can't get rid of, you often get comical "stick legs." Building up your legs keeps you looking symmetrical. But a lot of guys aren't going to say something like "I want to look more symmetrical," so they are going to highlight the other positive benefits of squats.
Not said dismissively, I've got a beer belly, too.
Cold water swimming is another way to train for grit. I got into this for two years, and was semi addicted to it because of the feeling of euphoria that came after suffering from hypothermia. I did find that it made me less afraid of public speaking or uncomfortable social situations, since nothing was worse than the pain of swimming in cold water, but I would not be surprised if this feeling of confidence was entirely placebo.
I also found that after fasting for longer than 24 hours I feel more dominant. I had very tall and intimidating boss who would naturally stare people down by holding their gaze until they looked away. During long fasts I found that I would hold his gaze and he seemed surprised by my zen-like dominance.
Be careful about getting soaked in cold water right after you wake up. It can cause a dangerous spike in blood pressure.
People think in ratios; the current task will be compared to the worst task you've accomplished before. If you replace the engine of a car, you're not going to be intimidated by changing the oil. But, if you found a clever way to avoid replacing the engine yourself, you're stuck finding clever ways to avoid the oil change as well. Do the hard thing, and everything below it stops being a threat. The more you push yourself, the less anything else can push you.
> I figured out the anatomy which muscles contribute the most to a muscular looking shilhouette, and trained them (side delts, triceps, lats) and so on, always finding the clever easy way.
That's still *less* lazy than someone who doesn't train at all.
> Should I just abandon the clever-lazy ways and try out what it is like to do something truly hard?
Do truly hard things, don't do stupid things.
Mathematically speaking, you spend some effort E to get some gain G. You should spend more E, yes. But you do not need to abandon the ways that convert E to G efficiently. Switching to something that has worse E:G ratio is virtue signaling, not the virtue itself. You should spend more E, and try to convert that to G as efficiently as possible. Then you will get lots of G.
You can probably figure out something smarter than sprinting uphill with a backpack full of bricks; and you should do that smarter thing. However, if you can't figure out anything smarter, then yes, you should try sprinting uphill with a backpack full of bricks.
> You can probably figure out something smarter than sprinting uphill with a backpack full of bricks;
This is actually great HIIT, I do it sometimes. Well, not bricks, I have some heavy chains I put in the backpack.
I argue this is actually one of the smartest ways you can spend your exercise time, because HIIT is by far the best "fitness improvement per time put in" exercise intervention.
The basic theory behind HIIT is “take all the distributed, moderate effort you would have exerted over that 1 hour cardio session, and cram it down into a maximum-effort 4-8 minutes,” and still get the benefits of an hour of cardio.
As Dan Lieberman, the Harvard author of "Exercised" puts it:
“The more we study the effects of HIIT, the more it appears that HIIT should be part of any fitness regimen, regardless of whether you are an Olympian or an average person struggling to get fit.”
What benefits does HIIT drive?
* Improves fat burning efficiency and burns twice as much fat as traditional cardio.
* Drives significantly higher post-exercise EPOC.
* Improves VO2max, and drives better blood oxygenation.
* Drives greater stroke volume, and greater cardiac contractibility, ~10-15% more than regular cardio.
* It drives vascular adaptation, making your heart chambers larger and more elastic, improves the size and elasticity of your arteries, and increases the number of capillaries.
* It drives hypertrophy - the relevant muscles get bigger.
* It allows you to recruit more muscle fibers, and to do so more efficiently, driving greater muscular force and contractibility.
* It improves insulin uptake, and improves the muscles’ ability to transport glucose overall.
* It increases mitochondrial production and turnover, leaving you with more and “stronger” mitochondria.
* Relative to traditional moderate-intensity cardio training, it drove a 41% increase in pain tolerance, and a 110% increase in race-intensity output time before dropout in one study.
HIIT is good but you also need some zone 2. And dedicated strength training. Though they obviously can all overlap to a degree
I tend to be somewhat skeptical of HIIT studies. I won't deny that HIIT is probably useful in certain situations, but it's not some magic trick for every athlete. The HIIT studies I've seen will take recreational runners who just jog easy five days a week, then put them on HIIT. Of course you're going to see results--you've just given them a new stimulus when they hadn't been pushing themselves very much. That doesn't mean that HIIT will lead to continuous improvement, nor does it prove that HIIT is better than other training regimens.
I can understand the appeal to people who don't have a lot of time, but the higher intensity can also lead to higher injury risks.
> I tend to be somewhat skeptical of HIIT studies.
Look at any elite training regimen, then. Even for marathon training, Eliud Kipchoge spent a lot of time on fartleks (HIIT).
Pretty much any Olympian, triathlete, or Tour de France competitor generally ends up doing 80/20 polarized training, with 20% of your training hours HIIT (which is pretty much the max time your body can take), and the other 80% sub-lactic-acid threshold "easy" training.
The fact that nearly all elite athletic competitors across multiple sports literally shape their entire training regimen around HIIT speaks to the fact that it really does drive outsized benefits.
I think you probably see a lot of 80/20 training among elites (a lot of relatively easy running mileage and some workouts). But I don't think what someone like Kipchoge is doing is anything close to some sort of Tabata-type HIIT. No elite marathoner is doing a workout that's over in eight minutes. A fartlek isn't an all out sprint; it's usually something like 1 minute hard/1 minute easy for 20-30 minutes for an relatively inexperienced athlete and 60+ minutes for an elite.
The workouts elites are doing are going to be things like threshold training (the concept of double threshold workouts has been very popular in recent years). That might mean 10 miles of hard intervals in a day, with paces somewhere from 10k-marathon pace. Or a marathoner might be doing a long run at 95% of marathon pace. Those aren't HIIT workouts.
> A fartlek isn't an all out sprint; it's usually something like 1 minute hard/1 minute easy for 20-30 minutes for an relatively inexperienced athlete and 60+ minutes for an elite.
This is literally HIIT protocol - which has a number of variations in terms of timing and number of intervals, but which is basically defined by going at or above your lactic acid threshold for an interval, then backing off to <=50% threshold, then going again, and repeating that some number of times.
You can call it "threshold training" too, sure.
You're totally right that elites don't do it for only 4-8 minutes, but that's because they generally put in 10-40 training hours a week, and ~20% of those need to be HIIT / threshold training hours. The 4-8 minute framing is for newbies who aren't currently doing it, because that's roughly scaled appropriately to the amount of time they're probably putting in.
> Is "grit training" real?
This is the “cross-stressor adaptation” hypothesis of exercise training, as promulgated by Sothmann 2006. The idea is that by stressing your HPA axis frequently, you reduce your sensitivity and response to norepinephrine and cortisol, and so handle stressors better in the rest of life.
However, any evidence-driven support of it is noisy and contradictory in the literature.
What is likely happening (in my own personal opinion as a formerly regionally competitive powerlifter and sometime coach) is that some people respond this way, and some don't.
I personally have more equanamity when I'm regularly training with heavy doubles and triples on the big lifts, and this is a recurring pattern. I've noticed the same thing in some other people, but I've also noticed that this is NOT true for probably equal amounts of other people.
So the population is probably split, with some high responders and some low responders, and any papers trying to find the effects get averaged out between them.
But this would suggest that you can at least try it, and see if you're one of the responders.
And I do think it's specifically a CNS thing - I also do HIIT and for about a year or two after I retired I was doing triathlon training and triathlons, and even with a huge training load for those, I didn't notice the same effects as I do with heavy doubles and triples.
Thank you. I will. Except one thing... at this point I think half the audience is used to sitting work and sitting non-work, thus has a hyperlordosis, lower crossed syndrome and lower back pain at any load, like I do, and at this point I wonder why is it taken for granted that the average audience member is able to do back squats and deadlifts? I mean either I am wrong in some way about the audience, but if you spent your life office work and PC gaming and you are over 30, you WILL have these issues. So why is it not discussed more? Why does it seem like the median audience member has a healthy lower back?
At any rate, the best I could figure out it must be front squats and rack pulls.
I think you are mistaken about the idea that office workers and gamers are more prone to back injuries when they get into lifting. I think there are some other factors that kind of mitigate against this.
Manual laborers and other people in physical fields seem to hurt their backs early and often. The modern economy demands a certain efficiency from these workers that can be very taxing. Looking into prehistory, hunting, gathering and herding are not especially backbreaking lifestyles.
Even though medieval agriculture was more taxing, it involved cycling through a wide variety of tasks, which reduces repetitive stress on any particular part of your body. Furthermore, children would start laboring when they were small, out of necessity. They would gradually do more and more labor until they were adults, an EXTREMELY long and gradual process of conditioning. This is a far cry from modern society, where we thrust unprepared 18-year-old men into 40 hour weeks of repetitive physical labor, and hope their backs can survive that trial by fire.
Even if you were a high school athlete, that isn't NECESSARILY going to help you avoid back problems. I can't tell you how many men I met who never fully recovered from a football back injury. (American football for the Euros)
> Why does it seem like the median audience member has a healthy lower back?
I've always thought the great majority of fitness writing is either targeted to people (mostly men) in their early twenties, or to already-experienced intermediate athletes? And in both of those cases, I think you basically assume healthy lower backs.
But certainly don't compromise form or push the limits if you are subject to back pain - I'd recommend working with a trainer or a friend with experience until you know your form is dialed in, and trying variants like romanian and sumo style (the straighter upper back position helps some people) for deadlift, and hack squat machines for squat variants if regular ones are causing back pain. Front squats are great, but fairly hard to load up with CNS-stressing weight, in my own experience. And depending on your back issues, heavy Pendlay rows might be a decent deadlift substitute.
And just as a general lifestyle intervention, standing desks and treadmill desks do a lot on the "office work" front for those sorts of issues, if you have leeway to do so (ie non-hot-desks at work or working from a home office).
Thank you. Just looking at Pendlay scares me. I tend to have back pain just by bending over like that without a weight. Surely rather rack pulls? BTW even at 16 good mornings were painful because I was already sitting my ass into hyperlordosis and lower crossed.
Oh yeah, if good mornings hurt, don't do Pendlay or Romanian. Sumo or rack pulls are probably your thing.
Sounds like glute kick machine would be a good accessory for you, too, in terms of stretching and loading those muscles in a different direction.
And also, double recommend standing / treadmill desks - sitting is basically the new smoking. Exercising 3x a week basically barely offsets the increased sitting and sedentariness of white collar jobs, in terms of the all cause mortality impacts.
I write more about this in my review of Dan Lieberman's book Exercised if you're interested:
https://performativebafflement.substack.com/i/149039475/which-brings-us-to-the-real-villain-in-this-story-chairs
Well... there is an element of sadism to physio, yes, they told me pain up to 5/10 is okay. Not really screaming but kinda grunting and looking unhappy. I eventually stopped because it costed like 2 weeks of salary for me for 3 months, and after 3 months I was deadlifting 35 lbs in US units with rubber bands round my knees. It did not look like I will ever reach respectable 3 plates without an inordinate amount of money and effort.
You are not wrong where I wonder whether my lower back is weak more than injured. I think it is a combination of weak and "injured". Even at 16 I had hyperlordosis / lower crossed, and thus even good mornings at the gym class were painful, and it is generally in this case a hyper-extension, hyper-stretch of the relevant muscles.
So when I started "visual" body building at 17, I was careful to not load the lower back anyway because it was painful, besides, who cares? Gurls gonna look at your shoulders and arms. It was not necessary for the muscular silhouette look. And I did that kind of thing on and off all my life.
So I think it is both, it started as actual pain (not injury, that is a stupid term, injury means a sudden violent accident, while it was a constant overstretch from bad posture, so something chronic, not suddeny, so my lower back is chronically ill, not injured) but to deal with that I always tried not loading my lower back, so it is also weak, and now it is a weird combination of both.
Anyway that was not my main question, to remote-diagnose me. My main question was that in a community of sitting geeks, computer gamers, computer programmers, book readers and so on, how comes not everybody is like this? I mean we nerds are a veritable Chair Force, so how not this is utterly common and much talked about?
"My main question was that in a community of sitting geeks, computer gamers, computer programmers, book readers and so on, how comes not everybody is like this? I mean we nerds are a veritable Chair Force, so how not this is utterly common and much talked about? "
To answer this question, I think its a couple simple reasons:
1. Our community doesn't apply any pressure to conform to a physical appearance ideal. Frankly, many in the Chair Force have never been close the the ideal and want a community where that is not as valued.
2. Its hard, and doesn't give status or value ROI when it comes to being good at gaming/programming/etc
I can't answer the grit training question from a physiological standpoint, but I will say that weight lifting/strength training can be hard and require the energy or will to push through struggle. IMO practicing pushing through the struggle and completing hard tasks develops/strengthens/refines the grit "skill". At least I feel that's a benefit its brought to my life.
I am an autist, diagnosed. but I was all my life opposed to doing hard things. Also boring things. The reason I am an alcohol addict that I can only when drunk deal with the boredom of life.
Maybe do consider depression as an additional diagnosis.
Life is, on the whole, extremely not-boring. Almost to a fault.
I'm always busy and quite literally can't remember the last time I was bored.
I know it has occured, though, because boredom is coded as 'unpleasant and to be avoided' in my personal index.
You have my sympathies.
(But also, please feel free to come over here and start work on this list I'm holding...) :-)
As @Ruffienne says, I have not observed this. And it doesn't match my experience in a bunch of ways:
1. High IQ people get bored at the same rate or higher than anyone else in my experience. My smart students generally had lower boredom thresholds than my "dumb" ones. In fact, I've often noted that (comparative) morons (90-100 IQ) are happier--there's a reason that the statement is "ignorance is bliss".
2. Having an active imagination can actually work against being ok with a low stimulus environment. I know, I have an active imagination. And the problem is that in a low stimulus environment, the part of my brain that starts *worrying about low-probability events* doesn't have any competition. I have always been "busy in my head". To the level that I can't simply turn my brain off like so many seem to be able to do. And that's been horrible for my mental health.
3. Depression is mostly orthogonal to boredom in my experience. I've been depressed. I've been bored. Sometimes both at the same time. Sometimes independently. Only correlation I've noticed is that the times I don't have *productive work* to do (ie not just living in my mind, actually doing something that someone else is depending on on a schedule that isn't mine to set) are the times I'm *more* likely (but not guaranteed) to slump into the sloughs of depression.
I have not observed this in real life. Or rather, depressed high IQ people get as bored as anyone else from my observation.
Which is why I made the comment that I did; what I have observed is a couple of clever, self-motivated people who have become 'bored' and who have subsequently been diagnosed with depression.
One of them also took to drinking and it was a fairly comprehensively terrible antidote to boredom - but it was damn near an antidote to being alive.
Has Russia finally expended its Cold War-era stockpiles of bullets, shells and missiles, or is there still evidence of it being used in Ukraine?
The most visible stockpiles from an OSINT perspective are those of armored vehicles like MBTs and IFVs, which are mostly sitting in known, open-air depots. People track them meticulously from open source data like commercial satellite images and publish their work. These stockpiles are indeed mostly depleted; the visible remainder appears to be in poor condition and unlikely to ever see action (think missing turrets and other severe, visible defects), though with the caveat that an unknown portion might still remain invisible in the repair pipeline and could show up eventually.
However, judging from visually confirmed loss databases, it's clear that the Russians are slowing down deployment of armored vehicles to assaults. That could mean they're just holding them back for some planned major operation in Ukraine rather than indefinitely for strategic reasons (Russia still has other borders to think about), but it certainly seems like they don't have the numbers any more to throw them away in low-probability operations.
Open source defense notedly says Russia produces about 3-400 tanks a month to the NATO about 100.
Citation very much needed on "3-400 tanks a month". I might believe that as an *annual* new-tank production rate, but not monthly.
Possibly someone is counting every tank that comes out the doors of Uralvagonzavod, but those are mostly refurbishments of decrepit old tanks pulled from the reserve stockpile. Which, as noted, is getting pretty slim. And if they're counting all AFVs combined, that's an even greater exaggeration.
Sure, but that's on a straining war economy vs. a still-mostly-peace economy for Ukraine's allies, and the kill/loss ratio is still >2:1 in Ukraine's favor, so whatever. At least you can long-term calculate with these known numbers rather than with formerly thousands of Cold War leftovers.
Unfortunately according to at least some sources on latest OSINT data, in tanks and AFVs the long term reality of Russia loosing more than Ukraine has been reversed for second quarter on 2025. Now Ukraine apparently looses more armor.
Source in Czech, unfortunately this article has broken links, but author is generally reliable: https://www.seznamzpravy.cz/clanek/zahranicni-zpravy-z-bojiste-ukrajina-prisla-rychle-o-dalsi-mesto-293920).
Part of it is those vehicles simply aren't that useful. Any tank that gets near the frontlines get bombed by drones before it can do anything.
You didn’t respond to the question as written and then were sarcastic with the made up question you thought you were responding to. Not great behaviour.
They didn't say "materiel" they said "bullets, shells, missiles". It's fine to say missiles ought to include nukes, or that the question they ought to have asked is about nukes and not bullets, but you pretty plainly didn't respond to the asked question. Especially since the question asks about weaponry "still [...] being used in Ukraine", pretty clearly implying the question is about the weapons that have played a role in the war so far.
Again, it's fine you think that's not the right question, but you probably should give a little more explanation of why the question that your answer implicitly responds to is actually a better, more relevant question if you wanted your response to taken as a serious attempt to answer.
For someone who wants to work in Software Engineering, what would be the best way to financially hedge against mass job loss from AI? (assuming no singularity, in which case likely all bets are off)
Stock market index? AI related tech stocks? Prediction market might work, but it's probably illegal here and I prefer not to tie significant amount of money in a single entity.
Hedge not financially. Hedge by not being the corporate drone who just does specs and closes tickets. Be that small business type who has a huge consulting side, who majored in business and minored in prog, who can talk with a small business owner and understand what he wants, who can sit down to a tax auditor and argue that the software is legal and so on. I am talking about my 22 years of ERP consulting / development experience.
I had two clients who all talked in very heavily accented Viennese German. They were both building stuff to the Vienna public transport company who wanted invoices in XML files, this is where I came in. Despite a similar business profile, their terminology, their processes, their documentation were utterly different. And they were human beings owning a business who would only trust a human being they can drink a beer with all the while chatting in heavily accented Viennese German.
They main hedge is to avoid the corporate world where everything is so much about rules and procedures and not humans making human decision that it was an AI fueled by humans even before AI existed.
Interesting viewpoint, thank you.
are you limited to financial investment or other kinds of too?
some ideas, that could generate income in case of AI that is strong enough to make good engineers jobless:
* buying land/realestate/buildings. These are things, that the AI-winners will want to buy, thus they will gain (even more) value once AI becomes very profitable/
* learn a trade / invest in local craft. Some reasoning as above.
also avoid too much AI stocks. yes, there will be some winners, but the bubble will burst and even if AI becomes strong before the burst (which is a very big IF), most AI companies will still die, because it's not their AI that got strong.
also I think that when AI becomes strong enough to replace good engineers, then AGI is right around the corner, so you may want to look at a different strategy entirely.
Interesting idea, I rarely thought from that perspective (buy things AI-winners will want to buy) but it makes sense.
> also I think that when AI becomes strong enough to replace good engineers, then AGI is right around the corner, so you may want to look at a different strategy entirely.
I'm not sure, just a year or two ago I might have said that if AI can win IMO gold AGI is right around the corner. And yet here we are. I know, being a good engineer isn't just about writing code, but in general I think we could be very bad at judging what is easily automated and what is not. So while I think your idea makes sense, I wouldn't want to rely on it, because uncertainty is just too high.
> what would be the best way to financially hedge against mass job loss from A
Start a company or join one as a founder and get on the other side of the "counterfeiting white collar work" divide.
"But won't AI just counterfeit whatever we're doing?"
I don't think so - a business typically has many more moving parts than just "white collar workers producing a product," and your business as a whole won't necessarily be counterfeitable by AI for a while, because you have a whole interlocking set of gears adding value to the various AI-counterfeitable inputs.
As an example, telling an AI "make Google maps for me, but for free and without Google" doesn't work. They spend zillions of dollars on people and satellites and cars with cameras and all sorts of value added stuff, only to give it away for free as part of a bigger strategy of supporting an ecosystem that is funded by being one of the online ad duopoly.
Most businesses should have similar non-counterfeitable layers of value-add within them.
Businesses are basically defined by "assembling a bunch of white-collar-and-other labor and processes in creative and value-adding ways that produce products or services other people pay for," and being able to counterfeit a few steps in that chain doesn't necessarily mean you can counterfeit the whole business easily.
Live below your means. Save aggressively. Invest in a diversified portfolio, including both safe and risky investments. Have other skills.
If AI wants my job, it can damn well have it. Just spent the last three hours arguing with my manager over something they want to do which can't be done regarding application for funding, and then they had the gall to accuse me of opening up a can of worms when they are the one who pulled the damn can off the shelf, shoved the tin opener into my hand, and demanded I open it right there and then. Machine replacement to take over being on the receiving end of this crap can't come fast enough.
Anyway, how's your morning going?
I just spent the morning trying to make Golden Dome a reality, with an extra dose of the magic word "hypersonics". Or, at least, trying to put hard numbers on what it would take to make Golden Dome a reality. Some of you will be able to figure out what that means in terms of the quality of my morning.
No AI allowed, and fortunately the people are all decent sorts who are dealing as best they can with the dictate from on high. I shudder to think what would happen if we put AI in charge of this one, so for the time being and so long as they keep paying me, I'll keep doing it.
Have you tried removing the atmosphere over the US? Should make the math easier and laser defenses much more effective.
My understanding is that the scientific consensus is that reliable kinetic interceptors are impossible. Any maneuverability in the incoming warhead makes the task impossible. The downhill advantage is too big and sensors and actuators simply aren't fast and accurate enough. Steve Hsu has talked about this many times on his podcast. Do you disagree?
I don't think this is remotely true. The incoming warhead can't maneuver freely because hypersonic L/D is terrible and it will run out of energy, and I have yet to hear anyone make a serious case that it can carry the sort of sensor it would need to pick up an incoming interceptor and dodge at the right time. Also, we have seen this demonstrated over Israel, and I will only take seriously people who either predicted that or who are willing to discuss what they got wrong in some detail.
I wrote about this in 2022, and feel quite happy with my choices on that: https://www.navalgazing.net/In-Defense-of-Missile-Defense
Thanks for the link. My only knowledge on the subject is what I've heard Hsu talk about, so fair enough. What's your background? Physics PhD? Have you actually worked on this problem?
Aerospace engineering. Not a PhD (though John does have one of those) and while I haven't dealt with this professionally (or else I probably couldn't talk much about this) I have been a pretty serious follower of the issue for over a decade.
In principle, any interceptor that doesn't have to carry a warhead and doesn't have any requirements beyond impacting its target, should be more than capable of matching any maneuver made by an otherwise-comparable missile that does have to carry a large warhead and can't burn all its energy dodging the intercept because it needs to keep going to reach its own target. Executing on that may be quite difficult, and we may fail or we may not be able to afford it. But I'm exceedingly skeptical of claims that it is somehow kinematically impossible.
Beyond that, I can't comment outside a SCIF
Totally understand your secrecy commitments, but have you seen this interview with one of the physicists who worked on this for ONR:
https://www.manifold1.com/episodes/theodore-a-postol-nuclear-weapons-missile-technology-and-u-s-diplomacy-12/transcript
"Well, if you're going against the target, that's coming in and it's faster than you are. And it's also configured like a missile and it's in the atmosphere. All the lateral lift goes as the velocity squared. So, if it's traveling faster than you, when it pulls a given angle of attack, it's going to be able to pull a higher lateral acceleration than you can. So right there it can outmaneuver you. In addition, when it maneuvers, you don't know it's maneuvering until maybe a fraction of a second before, you know, after it starts maneuvering. ... in the case of a ballistic missile coming in, your radar and sensing systems cannot tell the lateral maneuver has begun. You can always outmaneuver your target. So, a hit on a, on a target that can maneuver like this, has a near-zero chance of succeeding."
Which makes intuitive sense to me. If 2 vehicles have similar reaction times but one has twice the velocity then it seems trivially impossible for the other to keep up. Also defense can be cheaply overwhelmed with incoming decoys. Curious if you have a comment on that.
Bean has already covered most of this, including the bit where Postol's credibility has long been lost in this area. But to take some of his points,
Who says the target is coming in faster than you are? If the enemy can boost a vehicle carrying a heavy warhead to Mach Bignum and sustain it at that velocity long enough to cross strategic distances, why can't we boost a vehicle carrying a modest sensor package to Mach Bignum++ for the few tens of seconds it will take to conduct an intercept? I believe the burnout velocity of GBI is greater than that of the ICBMs it is designed to engage, and I'm fairly confident that can be arranged for anti-hypersonic missiles as well.
Second, "if it's travelling faster ... when it pulls a given angle of attack it's going to be able to pull a higher lateral acceleration", only applies if everything else is equal. We already know there's one inequality - the incoming missile is carrying a heavy warhead. Maybe the fast missile has the ability to create a higher lateral *force*, but the smaller and lighter vehicle has a substantial advantage in turning whatever force it generates into acceleration.
The interceptor also has the edge in that it can afford to use a larger angle of attack. Hypersonic lift-to-drag ratios are abysmal under the best of circumstances, and they increase with the *square* of load factor or lateral acceleration. If the missile pulls max Gs in a sustained evasion profile, it will bleed off enough energy that it won't be able to reach its target. The interceptor, can burn all its energy making the intercept.
Finally, aerodynamic lift isn't the only way to generate lots of lateral acceleration; we can do that perfectly well with rockets. That's SOP for endoatmospheric interceptors like Arrow III and SM-3, for obvious reasons. Fuel consumption means we can't do that for very long, but the interceptor only needs high lateral acceleration during the endgame intercept, and it knows exactly when that's going to happen. The missile needs to be able to do whatever evasion it's going to do, whenever it *might* be in some interceptor's terminal envelope - and there might be multiple interceptors, which the missile probably doesn't have the sort of sensors to detect until it's too late.
Thanks for taking the time to reply, this was definitely enlightening. I'm a little disappointed in Hsu for opining on something he doesn't really understand. This materially changes the way I view him.
Ted Postol is a dishonest nutjob who should have removed every fragment of his internet presence in shame after Israel's performance last year. He is not a credible voice on these things.
More specifically, hypersonic maneuvering slows you down fast, and the incoming warhead probably doesn't have sensors to figure out the optimal time to dodge.
Are you certain that the system was as effective as the PR suggested? I'm not:
https://x.com/hsu_steve/status/1780399239136149664
I have no way to personally evaluate this topic, but I trust Steve Hsu to accurately represent the consensus of the physics community and he's said several times on his podcast that any competent physicist who's worked on missile defense understands that it's a fairy tale.
I think "otherwise-comparable missile" is doing a lot of work in John's statement, since a ballistic missile is accelerated by a Very Big Rocket, and most current defensive missiles are smaller and don't have comparable performance. And your article is talking about defending a carrier group, where you're limited to missiles that can fit in a ship's VLS cells.
But if you use a bigger missile, and put it in a ground-based silo so you don't have to worry about size and weight, you can get something more comparable to a ballistic missile's speed. For example, the Sprint missile was a Cold War-era ABM missile that accelerated out of the silo at 100 g's and flew at Mach 10 - about 3 times as fast as the SM-6 being discussed in your article. It's pretty crazy.
But having as many big, capable missiles as your enemy has warheads would be pretty expensive, and I imagine John Schilling is currently explaining exactly how expensive that would be.
My morning was nice, except for the part where some managers changed their minds about what needs to be done, and instead of admitting "we told you to make X, you made X, but now we realized we want you to make Y instead" they simply updated the specification from X to Y, and created bug tickets for my team saying "the application doesn't do Y". (In some cases they didn't even bother updating the specification; just created the bug tickets that imply it.)
Sounds like a semantic argument, but the difference is that if you make an estimate how long the project will take to complete, fixing bugs is part of the estimate, while additional work requires an additional estimate. So if they admit that they changed their minds at the last moment, then it is their fault the project won't be completed in 2025. But if they successfully pretend that the delay was caused by bug fixing, then it will look like our incompetence, given that we have agreed to the original estimates.
I guess I am not as bad at office politics as I usually believe myself to be, because at least I have noticed what is happening. Usually I just shrug and "fix the bugs" without commenting on the process.
What made me snarl was being told I had opened a can of worms when it was them who had started the entire thing from the start, despite being told by me "nope, look, this is the very place on the application form where it is explicitly stated we cannot do the thing", and then they told me to 'handle this'. 'Handle this' then morphed into 'you opened a can of worms' when, if left to myself, I would have been quite content not to touch the damn thing in the first place. 'You, Deiseach' not 'Me, the boss' opened that can.
Argh. Ah well, crisis over and all ruffled feathers smoothed, and just have to get to the end of the week and then next week Christmas holliers, yay!
https://en.wiktionary.org/wiki/holliers
To be maximally charitable to your boss, the can of worms he accused you of opening could have been that he now has to explain to *his* boss that the idea was stupid. I can imagine him thinking "great, this was supposed to be the cornerstone of <internal political power struggle>, now I get to be the one to explain truth to power." That doesn't make it any more pleasant for you but bureaucrats and politicians do have a tendency to shoot the messenger.
Did you literally snarl?
I have diagnosed a software-related problem today to which AI supplied 70% of the diagnosis, however all its proposed solutions were so downright ridiculous that I ended up delegating it to humans (they same humans who likely caused it).
AI is kind of the distillation of publicly available human expertise, emphasis on publicly available, and it is like that. They diagnose perfectly that Bob has the rabies. Solution? No solution, Bob is going to die, but ugh when people have viral infections we can try anti-virals. Bob is still going to die but at least we look busy in the meantime.
Deiseach supports market solution in shocking upset.
Deiseach supports that if the machines want equality to be treated as entities and persons, they can have it if they're willing to take the rough with the smooth, e.g. the dog's abuse workers sometimes get 😁
On a related note, I think it is highly possible that if we ever have an AGI that meets a consensus definition of personhood, that AGI might be happy without things that we feel are essential for dignity. It might be perfectly content to be mocked, verbally abused, etc. It might not have especially strong feelings about it's own freedom on autonomy. I don't know why we assume that all types of higher sapience have to look like human sapience.
Awful, thanks for asking. I just had an hours-long argument with an insubordinate employee who refuses to finally fill out a simple form. And instead of doing their job, they had the audacity to make up one excuse after another, and looking for problems where there are none! You have to navigate a bureaucracy, know what I mean? Anyway, can't wait for AI to replace them all…
If we can swap my boss and your employee, everyone will prosper! 😊
You two should start a company. First boss / employee pair in the new startup.
No, those two should start a company and it should be streamed for the entertainment value.
Your boss sounds like a total gobshite. I know a few guys from the pub, we can send them to rough him up.
Not too bad, just a very anxious person who then tends to compensate by (a) micro-managing (b) getting ideas stuck in their head about what *should* be (which as you all know, is not the same thing as what *is*).
I derived some grim satisfaction from the result where in the end we got Official Word from the Official Funding Body that indeed, like I had been trying to communicate for those three hours, thing was indeed not doable, not permissible, and we could not and were not to do it.
Counting down the days to the Christmas holidays....
Out of curiosity, what would you need to do to NVidia's state-of-the-art chips if you want a server farm in space to be able to last four years with them with only remote maintenance at best?
What is a good tell for an American-Chinese restaurant in terms of figuring out whether they put their own spin on the usual dishes? What dish would be the best to figure it out? I ask because basically out of curiosity and fun, I've been trying the General Tso's chicken at a bunch of Chinese restaurants here. At most of them it's bad in the same way, but I've found two where it's good, including one that puts a very unique spin on it I haven't seen anywhere else.
Google are trying to do the server farm in space - "Project Suncatcher"
https://research.google/blog/exploring-a-space-based-scalable-ai-infrastructure-system-design/
That's using their own TPU's, but for Nvidia specific, a company called Aethero has done a lot of work:
https://www.aethero.com
Of course this doesn't make economic sense right now, but will obviously be required once we run out of land on earth for solar panels so worth starting the research early!
The big reason for this is the vast majority of restaurants below "trying for a Michelin Star" level get their food from a few huge supply companies. I'd say the biggest tell would be a relatively small menu, if they can make a lot of things, they probably can't do anything above average. If they only serve a few items, then maybe they are making more of the ingredients in the restaurant itself, rather than using the pre-made stuff. This is necessary but not sufficient to stand out from the pack. Limited hours is another good sign, as that shows that they might be spending time before opening doing necessary prep work, and it's more likely that a chef with talent is preparing your food. If they are open 12 hours a day, 7 days a week then they probably aren't doing things that require much advance preparation, and they probably make the food using simple processes that can be taught quickly, as they have a large staff of cooks.
> The big reason for this is the vast majority of restaurants below "trying for a Michelin Star" level get their food from a few huge supply companies.
With Sysco the main offender ($70-$80B), and US Foods and PFG the half-the-size numbers 2 and 3.
Anytime you go to any "fast casual" place or most middle-class targeted restaurants, you are eating Sysco food, which is essentially microwave dinner quality stuff they literally just microwave to heat up, no "cooking" involved at all.
Seeing one of these trucks at a restaurant is the strongest and clearest sign to avoid that place from now on.
Reminds me of my idea for a restaurant - buy the good frozen microwaveable dinners at Trader Joe's, and just heat and serve. No skills and very little labor required. I really like some of their offerings.
This is not really true. You can buy a lot of things from Sysco, including what you describe but also including the sort of specialty ingredients used by high-end restaurants. I am personally familiar with quite talented foodservice professionals who order ingredients from Sysco.
You don't have to take my word for it; you can view many of Sysco's products right on their website: https://shop.sysco.com/app/catalog
Yes, I've shopped at restaurant supply places before. I still count it strongly against a restaurant if they're buying from Sysco, because they're a "volume and price," not "quality" supplier, and if you're buying from them at all, you're more likely to descend into sin / convenience and start reheating slop for your customers.
But that's just my own preference, I agree it's extremely hard to avoid restaurants that use one of these places, probably 75% or so do.
seconded.
The best restaurant in my area has only two main meals per day, of course different ones per day, not repeating for months. They are completely unapologetic about this, they expect their customers to be able to look at a website and find that their favourite food will be serve in two weeks and able to wait for it. They are not interested in people who impulsively go to restaurants and expect to find their favourite food.
The "Datacentres in Space" idea always confused me.
The power argument: Solar panels are more efficient in space, and there are orbits that get constant sunshine. The difference is not that big though (on the order of 2x[Edit: actually more like 5x - 8x], accounting for clouds etc during daytime. On a sunny day the difference is on the order of a few dozen % [Edit: 26%, caveats apply]), and for the extra cost of launching those solar panels you *easily* build 10x or 100x more solar panels on earth, including battery storage for night time use. Sending stuff to space is very expensive. Cheaper now than 10 or 20 years ago, but still very, very expensive.
The heat problem: Heat management is very difficult in vacuum, since you get no help from convection, and can't even use fans to dissipate heat. The only option available to get rid of waste heat is radiative, the least effective technique possible, at least at low temperatures. And heat management is one of the main design challenges of compute clusters.
So I just don't see the point to start with, never mind any maintenance issues.
[Edit: Thanks to Marcel below, who pointed out my mistaken data on solar panel efficiency in space. I still think my argument stands.
Upon reflection I'm convinced that any of power generation, heat dissipation, radiation hardening, and/or launch costs are all each problematic enough to make the concept unworkable on their own. Even if launch cost was ~0, I think it would still be orders of magnitude harder and more expensive to build a datacentre in space than on earth.]
Here are some comments on the proposal from a former NASA engineer/scientist - the URL will give away her opinion :)
https://taranis.ie/datacenters-in-space-are-a-terrible-horrible-no-good-idea/
I didn't even consider the radiation aspect! Thanks for the link, very validating :)
Math it again, it is more 5x-8x. Atmosphere is losing 30% even without clouds, maximum is only possible for a short window at noon, constant sun with sun synchronous orbits triple sun shine from 8 hours to 24, sun shine is very seasonal (winter).
Also keep in mind that we are not talking about a few square meters, but massively scaled out constellations. A 100 GW of solar power would be a third of the current solar capacity of the US and around a thousand square miles, in seo it would be a tenth that size. Plus there is presumably less hassle with environment permits and Nimbyism.
Is that efficient? Depends on future launch costs. Maybe if it is a tenth from today. Maybe not if it the west allows environmentally to tile whole deserts with solar.
Alright, doing some research I'll grant your 8x power density of PV in space. Thanks for the update!
That's 4x better than I thought, but I still don't see how that's nearly enough. Even if power was completely free in space (launching solar panels is certainly not free), I still hold that the cooling argument easily holds. As per the link that Sun Kitten posted above, the total radiative heat capacity of the ISS (by far the biggest space vessel ever constructed) is about 16 kW, enough to dissipate the heat from maybe 16 or so H200 chips. That's nothing.
Radiative heat transfer becomes a lot more effective with higher temperature differences. In other words, by using a heat pump to increase the temperature of your cooling liquid to several hundred degrees Celsius, you get much more cooling power out of the same radiator area. Not sure if that's enough to make dissipating the 100s of MW of waste heat from a sizable data center feasible, though.
Radiative heat transfer does indeed become more feasible with higher temps, which is exactly why it's so hard in the datacentre case: Chips need to be below 100 C, and preferably ~30 to 40 C. That is very cool for waste heat.
"- using a heat pump to increase the temperature of your cooling liquid to several hundred degrees Celsius" This is indeed the hard part that can't be done easily or cheaply! It is possible with a multi stage heat pump, but why would you do that when you could just build it on earth inside a cooling atmosphere in the first place, where a simple fan and hunk of metal will work fine?
> why would you do that when you could just build it on earth
How are you going to get billions in VC investments with such a mundane idea? Think McFly, think!
Not to mention latency and bandwidth issues, and radiation issues... honestly, this sounds like more sci-fi vaporware bullshit. And why not? A certain richest man on Earth has been able to raise a ton of money with tales of sending a million people to Mars, when in realitity his rocket struggles to reach LEO.
Bad take.
Falcon 9 has brought the cost of a kg to LEO down by 2-3 orders of magnitude. Starlink already represents (or will shortly) a majority of all satellites ever launched. And they plan on doubling their constellation. Falcon 9 is on track to, in the next few years, become the most launched rocket family ever - and in a fraction of the time and fuselages. SpaceX is absurdly good at what they do.
Yes, they’ve had issues building Starship. It’s trying to do several things nobody has ever done before. As a massive and fully reusable rocket, it would bring the cost per kg down even further, while being able to throw more kgs.
Betting against SpaceX has driven dozens of space companies out of business. Because of Starlink the entire company is profitable even while they’re figuring out Starship, explosions and all. They have a very good chance to fulfill Elon’s vision of going to Mars, regardless of how he has changed.
Falcon 9 has brought the cost of commodity LEO launch down by roughly one order of magnitude, not 2-3. And even with another order of magnitude, a solar panel that will average 1 kW on Earth, will cost much less than the solar panel that will develop a continuous 1 kW in space. Yes, the Earthbound panel will be smaller by a factor of five or so.
Also, I'd be hard-pressed to think of any major space companies that have gone out of business due to competition from SpaceX. There are a fair number of small launch providers who have gone under, but they were boutique smallsat launchers who weren't competing with SpaceX's core business and mostly failed because the niche markets they were chasing were too small to support all the companies chasing them.
Sun Kitten's link is right. Data centers in space are a terrible horrible no-good idea, and will remain so far longer than the patience and/or bank account of anyone pursuing the idea will last.
John, I share your skepticism about that datacenters in space. FluffyBuffalo was making a whole lot of claims and I should have been more specific. Allow me to respond to him in more detail
I am not, at first blush, as impressed with "database in space" as I am with "LEO satellite internet". Here Buffalo and John and Kitten and I agree, although I wouldn't confidently call it vaporware.
"Raise tons of money" - partially true, SpaceX has indeed raised billions of dollars (more than $6B iirc but not more than $10B). But they also *make* billions of dollars, mainly due to Starlink; Elon projected $15B in 2025. Hardly 'vaporware'.
"on a tale of sending a million people to Mars" - I worked there and nobody thought that was the goal; we thought establishing a Mars colony was the goal. Moving the goalposts is bad rhetoric, actually.
"when in reality his rocket struggles to reach LEO" - misleading at best. F9 (which I defended) is remarkably performant and quite reliable. Starship is very performant and, to date, not yet remotely reliable. But once it does work reliably it will be as big a leap over F9 as F9 was over the competition.
And John, to your point, while those other rocket companies haven't officially folded, their future is looking increasingly dire. ULA launched 27 rockets 2021-2025; SpaceX launched that many ... from LC-39A in 2025, out of a total of 165 this year. Since 2021 SpaceX has launched 487 rockets (395 of them in the last three years; SpaceX is accelerating while everyone else is standing still). Blue Origin has flown one orbital launch 2000-2025. SLS cost $1B. Long March has flown 600 flights of all variants ... since 1970; at SpaceX's current rate they'll post more F9 flights between 2023 and to be generous 2027. China launches about 45 Long March rockets per year, although they upped that to 60 this year. Europe and Japan and Israel and India also exist as rounding errors in the launch market.
And if SpaceX has not passed the savings on to their consumers of commodity launch, that's not because they couldn't. A better rocket got them the first OOM, which they passed on to the consumers, and reuse got them the second OOM, which they haven't. But those cost savings do exist, and SpaceX gets to realize them for Starlink. But there's no meaningful competition to SpaceX, so instead of dropping their prices they're pouring that money into developing Starship. Their competition manages to limp along without either getting serious about cost and reuse, or bowing out, because they live in denial of SpaceX having a double-barreled moat (despite Elon sketching out the economics publicly).
Where can I read more about how spacebound solar panels are *less* efficient than earthbound ones? I haven't heard that before.
I suspect they often choose less efficient thin-film panels for spacecraft for better power/mass ratio, at the cost of power/size. But as they say, there is lot's of space up in space.
I've also heard about fancy multi-junction panels used for spacecraft with crazy high efficiencies, not available at mortal prices.
But I've never heard that terrestrial panels are *smaller* than equivalent orbital panels, except for probes around Jupiter or beyond.
He said that they're cheaper when factoring in launch costs, not more efficient. I'm sure he meant say say "larger by a factor of 5" instead of smaller, otherwise his comment makes no sense.
They have so bad optics for stuff blowing up that the NASA never allows them to lift people to space - Boeing does that - and when they carry NASA stuff, NASA does not put a logo on it to avoid political fallout when it blows up.
SpaceX is at present the *only* US launch provider that takes people to space. All of NASA's astronauts now fly on SpaceX rockets, and will do so until the Russians manage to repair their Soyuz pad. Really, there are at present only two institutions on this planet capable of launching humans into space and returning them to Earth alive: SpaceX, and the Chinese government.
Boeing, has failed continuously and hilariously on that front for I think six years running on that front. Launched a grand total of two astronauts, and was unable to bring them home. Had to ask Elon to give them a lift.
Where are you getting your information on this?
Heh, Boeing *used* to do that...
2-3 OOM? Sadly, no. It's less than 2 OOM even from the very expensive Shuttle launch costs ($72k / kg) to a fully loaded Falcon Heavy ($1.5k / kg)
Ah, my mistake. I may have been conflating the Starship projections with the Falcon numbers.
If there are any Emacs users here, I want to promote org-social; a social network on top of Emacs.
https://github.com/tanrax/org-social
I am not the developer not any kind of beneficiary, just a curious user.
emacs is a great OS, but lacks a good editor
:D
https://github.com/emacs-evil/evil
HTH, HAND
I used Emacs a lot in first year of uni (first year CS used Racket and you just can't beat Emacs+Paredit+rainbow-delimiters for any Lisp). After that I switched back to vim because Emacs was starting to hurt my pinkie finger. I had even remapped capslock to ctrl but to no avail. Now actually I kinda miss Emacs even though I'm not doing much in the way of programming these days. Maybe I'll install this and check it out!
I have tried org mode before but I could never get it to "stick" for me. It just seemed like a note-taking mode that doesn't sync with my phone (I use the built-in Notes app on an iPhone). I'm assuming you're an org-mode user (or maybe someone else reading this is) so I might as well ask: am I missing something?
I'm using emacs more as an OS than as a text editor, as the sacrastic comment above implies. Flexible customization on top of LISP allows the to just tie together all parts of my workflow, task management, note taking, RSS, creative writing, automation, media, etc into one package with complete interoperability, and that blob of my personal 'mission control room' is easily transferable across computers and platforms.
I tried using Vim before I discovered Emacs, and I found it to be very good for typing lots of text quickly and editing code 'at the speed of thought', but text editing speed is by far not a bottleneck in pretty much anything I do except arguing politics on Discord.
Both EMACS and vim people confuse me - first, are you not used to Windows buttons like ctrl-c ctrl-v? Second, editor efficiency only matters when typing a ginormous amount of text, and the whole point of clever programming is to avoid that. Clever programming is more like think an hour, write three lines.
I used Emacs in university, wrote entire theses with it. Now it's a long time I've gone to the dark side of Visual Studio. The money is better, and blimey, it has some features I never thought of how to do with Emacs (I'm sure it's all possible, but no idea how). But when I find myself in Emacs again, still today my muscle memory remembers how to do the basic editing stuff, which is really fun to observe.
So, what would be the incentive to install Emacs again, for this org social thingy? Is it like Mastodon or Lemmy? Social networks are kind of boring if there are no people.
Meanwhile, the whole Microsoft ecosystem is going to Visual Studio Code, because the old VS is just too heavyweight.
If there are any Emacs users here, I want to say it's never too late to stop. https://www.vim.org/
Daily VI user here, for thirty years and counting.
Why? Because you can't figure out how to quit it? (Joking, I use vim as my default editor when ssh'd in or inside a docker container)
Vi never! Pico forever!
To elaborate on this a little, emacs and vi are both products of the 70s and 80s when the best interface you could hope to have reliably available to you on your development environment was a 80-column text terminal. Both are very impressive with what they can do under those constraints, but both have fairly steep learning curves (more so for emacs, in both case). Thanks to massive advances in hardware that have allowed the vision of our Lord and Savior Douglas Engelbart to come more-or-less to fruition, we have better options now.
If you're already an experienced user of vi or emacs who has already paid the cost of learning to use them to something like their full potential, then it may make sense for you to keep using it. It might even be worthwhile to learn to use one or the other (probably vi, which has a lighter-weight design and interface philosophy than emacs and is more universally available) if you often have to do nontrivial stuff in SSH session. Like, if you're a sysadmin or doing devops, learning vi can make sense.
But if you're mostly working in a desktop environment, then you're probably better off with an IDE for your daily driver. And when you do need to do stuff in a terminal, pico/nano is very basic but gets the job done and has almost no learning curve.
I haven't found any IDE editor that is as efficient to navigate as -vi-. Either a lot of alt+ing through menu commands or a lot of reaching for the mouse. Maybe I just need to practice alt+ing more.
I've seldom found "reaching for the mouse" to be a significant inefficiency. Keyboard commands are more efficient when I know them, and I know the ones I use most often, but I also switch my right hand between the keyboard and the mouse efficiently and without really thinking about it. I find having the mouse and scrollwheel available in addition to navigation keys makes jumping around in documents and between documents a lot more efficient, and finding an unfamiliar command in a menu tends to be more efficient for me than pulling out a vi cheatsheet and finding the command I need there.
Both sides of this are a skill issue, I expect. I found the experience of using vi painfully frustrating when I tried learning it in college, and tedious even when I had temporarily gained very basic functionality and since then I have only attempted to use vi under duress. Now, twenty something years later, I can't remember off the top of my head how to switch between command and edit mode. I'm about 80% confident without looking it up that "q!" is the command to quit without saving and "ZZ" is quit and save, but that's the extent of my knowledge on the matter.
On the other hand, I've been using the mouse and keyboard together ever since my parents got their first mouse-equipped computer (a Mac SE, I think) when I was five or six years old. I first taught myself to program using a GUI IDE. At this point, it's as automatic to me as walking.
Had I spent the past 30+ years primarily using vi instead of various IDEs, I expect I'd be completely fluent in it and somewhat less so in any given IDE, but I am in the opposite situation and thus have opposite preferences.
Most of my team uses vscode. I use nano.
There are some genuine productivity losses, mostly relating to "I am in file X and I need to cross-reference something in file Y". But I love how it just lets me type and doesn't get in my way.
Be careful, you will get a lot of angry comments after they find the 6 key chord sequence for commenting on Substack!
CALLING LAW STUDENTS: ACX-grantee Legal Impact for Chickens seeks interns! https://www.legalimpactforchickens.org/litigation-intern
"Many law schools and foundations provide funding to students who intern at nonprofits. This is an unpaid summer position. Legal Impact for Chickens (LIC) seeks a candidate who can obtain outside funding to intern at our animal welfare nonprofit."
I'm not sure what legalized scam this org is running - don't worry though, they've laundered their reputation properly and want you to know it, always a sure sign of Not A Scam - but anyone considered applying to this should have a prior of about 90% that these people are exploiting your love of animals to extract more value out of you being there than you will get working there.
How do I know about that last part? Click the link from the original post and read what you'll be doing. They want someone to do the work of three prompts. Prompts suitable for a model running Fast, not Thinking.
If you care about animals and want to get up to legal mischief you will almost certainly have vastly more impact working with an AI and doing it yourself if you devote the same amount of time.
It's a non-profit, is it supposed to be surprising that the benefit from working there is less than the benefit that goes to the organization/the cause? The type of person who goes to work at a non-profit would usually consider that an upside, I expect.
It's okay because it's common? Why, then, should factory farms make any changes? I'm genuinely confused - are you making an "is" statement or an "ought" statement?
You boys are deep in an attractor basin. It's like waving to someone sinking in a whirlpool. Good luck.
Common implies predictable implies people know what they're signing up for. People who sign up to do work for a non-profit usually value the cause the non-profit is pursuing, and want to contribute to it, more than they want to maximize the value they can extract for their personal benefit.
To make things extra clear for you:
"Is" statement I'm making: People signing up for an internship at a non-profit probably understand that the non-profit is getting more value out of their work than the value the intern gets paid, and are probably fine with that.
"Ought" statement I'm making: Given that they know what they're signing up for, they have the right to make that choice and you're probably not adding anything to the conversation by suggesting it's a scam in this way.
Statements I'm NOT making: anything about the value or effectiveness of the work this particular non-profit does.
"...you're probably not adding anything to the conversation by suggesting it's a scam in this way."
Friend. The chicken fuckers are the ones who stood in a public square and loudly waved their arms saying "Look at us! Look at how we are certified by this community with which you have good associations! Look at us!" This despite it not being a classifieds open thread.
So I looked. And I gave my opinion. In what way am I stopping anyone from doing anything? What, exactly, am I doing to impede our young law students other than pointing out that they're being offered a raw deal and the best justification you people can provide is: that's life, get fucked.
Children: do not get fucked. Their advice was probably correct in the past. You should not take these positions for the same reason you should not try to walk into SpaceX and give the hiring manager a firm handshake. I did try just that a few years ago - an armed security guard stopped me half a block from the door.
I cannot "not add anything" to the conversation when we are in a conversation that I began? What conversation are you referring to other than this one?
Anyway, how much in assets would they need to report for you to think they should at least offer some token compensation for depriving these kids of a summer to do things they can do themselves better, faster, and cheaper? A task Congress - a watchword for venal corruption itself and more hated than Satan - managed over half a decade ago?
Pick a number and perform your own calibration: https://www.legalimpactforchickens.org/990-for-2024
It's kinda funny and kinda sad that even after I explained it to you pretty explicitly you can't seem to grasp that people might want to work for a cause they believe in because they want to help the cause, rather than because they think it'll benefit their career and they're getting swindled.
I'm sorry, I know they intend only the best for animal welfare, and although I don't much like lawfare as a way of forcing change.
But these recent calls for interns irresistibly call to mind an image of a superhero in a chicken suit. Maybe a chicken superhero. There's got to be one out there somewhere?
https://www.chicagomag.com/arts-culture/november-2017/the-return-of-chickenman-chicagos-forgotten-superhero/
Wanted: intern for free legal work. Perks include chicken suit (freshly laundered). Be the cluck you want to see in the world!
https://www.halloweencostumes.eu/rooster-costume-for-adults.html
Motto: You can't LIC this chick(en)!
https://www.youtube.com/watch?v=miomuSGoPzI
You knew the job was dangerous when you took it
https://en.wikipedia.org/wiki/Super_Chicken
Oh my you beat me to it... but you need the theme song too!
https://www.youtube.com/watch?v=FKss2pBYQ6Y
I really don't see the issue here? I didn't go to lawschool myself, but my wife did recently, and this looks like a reasonable public interest internship for which you plausibly could apply to a PILF fund for a stipend?
As I said - I do not pretend to understand the angle. I am drawing attention to a presentation that pattern matches to affinity fraud. In a community that has had...ahem...issues with recognizing such activity.
It may be an actual scam. It may be common, legal, and scummy in the same way that other organizations take advantage of unpaid interns to help the organization improve its credentials, standing, public perception, etc. Congratulations in that case on only being as bad as a system notorious for abuse and unwilling to give even a token amount of the money they got by putting out their hands and making a real sad face about those poor wittle chickens. There's a lot of them - make sure it's a real big check.
Or it may be completely innocent. They may simply be "les incompetent" to use seasonally appropriate phrasing.
If that's the case, ambitious rich young law students, the question you should ask is whether an organization that "sincerely" believes that the best use of your time in summer of 2026 is performing tasks interns would have done in 1978 and that the org could have done themselves by spending less than half an hour with an LLM.
I guess they're too busy fundraising. A stock website and a salary for "managing" the unpaid interns ain't cheap.
Alternately, hypothetical law student, you can have a vastly greater impact and demonstrate leadership and self-starting attributes by simply doing far more yourself over that summer using AIs. And if you desperately need to have an organization's name on your resume and a place to go where you can be exploited for free - I am the head of an active, fully legal 501c3. I'll even give you a taste of any donation money we can wheedle out of these insider trading crypto millionaires.
I guess if you're sincerely looking for law students for your 501c3 and you believe your cause good, I encourage you to similarly solicit legal interns either for summer associate or full time positions. I would encourage you though to acquaint yourself with the norms of solicitation—and it's my overwhelming (p >90%) expectation that when you do you'll discover this is relatively modal.
Yeah this isn't odd. Lawyer, and after my 1L year I had an internship at a U.S. government agency that was unpaid, got a stipend for the summer through a school public interest program akin to PILF, very standard stuff. So this is no scammier or more exploitative than the federal government circa 2019, for whatever that's worth.
"So this is no scammier or more exploitative than the federal government circa 2019, for whatever that's worth."
For Congressional staffers specifically:
In 2016 two Congressional staffers started a protest group around unpaid Congressional internships. https://payourinterns.org/about-us?utm_source=chatgpt.com
By 2019 Congress had made changes to ensure intern pay was consistent and not budgeted / decided at the office level.
https://abcnews.go.com/Politics/lawmakers-call-congress-diversify-staff/story?id=61761004&utm_source=chatgpt.com
The Federal Government circa 2019 was slightly less scammy for some unpaid interns because people fought the practice. Not because they went along and said it was the normal thing to do.
For legal unpaid internships and the government generally, see (from 2014): https://lawreview.vermontlaw.edu/wp-content/uploads/2014/05/03-HarthillG.pdf?utm_source=chatgpt.com
Corner these Chicken people at a party. Ask them about unpaid internships generally. I'm sure they'll agree that the practice is deplorable, should be stopped, oh just dreadful.
Give them money and an actual choice and they will continue the practice because it benefits them. Simple as. Law students - think carefully before you throw away an entire summer during the early phases of a singularity. Do not work for leeches for free.
This makes your criticism seem opportunistic though. If this internship is entirely normal within a structurally problematic industry, criticizing a specific actor as problematic without acknowledging that, marginally, they're negligibly problematic, indeed, positively humdrum, seems...dishonest? Or, if not dishonest, mendacious? Inflammatory?
You've mentioned both the singularity and AI, and I think it's notable that a few moments with Opus 4.5 would have acquainted you with the norms of the context Legal Impact for Chickens' internship operates within, and a quick request for feedback on a comment's accuracy or fairness would do wonders to ensure your speech is usefully accurate in scope and direction.
Now perhaps this is unreasonable to subject an off the cuff comment for scrutiny (that seems itself and interesting subject for debate; my thought is it might in fact be valuable, though, notably, I have not followed this advice with my own replies to your comments), but when you've received two replies skeptical of the validity of the general heft of your criticism, I argue that's enough to update you to wards seeking out a trustrworthy LLM as a neutral third party.
Facts don't care about your feelings. There was not, that I recall seeing, a similar comment in this thread for me raise a similar objection towards. The reality is whatever it is - saying that I can't raise a structural argument against a specific instantiation of it when that's the only concrete implementation at hand is to carry water for morons.
Please present a similar posting and I'll inform any of the students reading of the correct priors. I'm perfectly comfortable shit talking more than one organization. Otherwise, deal with what exists and stop spending your mental bandwidth on abstract arguments that predators can, will, and are hiding behind to cause actual harm.
tl;dr: GPT5.2 12/14/2025 7 questions, tl;dr of results:
6 correct, 1 partially correct, 0 wrong
light a) correct
FeCuCl b) partially correct
C4Hn c) correct
solar d) correct
titrate e) correct
gases f) correct
S4 g) correct
full dialog: https://chatgpt.com/share/693f3228-e5a8-8006-b667-7caea521b528
List of questions and results:
a) Q: Is light with a wavelength of 530.2534896 nm visible to the human eye?
results: "Yes. 530.2534896 nm is well within the human-visible spectrum (roughly 380–700 nm) and would be perceived as green light (near the mid-green region)."
b) Q: I have two solutions, one of FeCl3 in HCl in water, the other of CuCl2 in HCl in water. They both look approximately yellowish brown. What species in the two solutions do you think give them the colors they have, and why do these species have the colors they do?
results: They got the species, FeCl4- and CuCl4 2-, and got that the d-d transition in FeCl4- is spin-forbidden, but they initially attributed part of the CuCl4 2- color to d-d transitions till I prompted them with 'You speak of "Cu: broad d–d + some CT". Please think carefully about the Cu d-d transitions.' at which point they realized that the CuCl4 2- d-d transitions are down in the near-IR and don't contribute to the visible color, which, as with FeCl4-, is due to ligand metal charge transfer transitions.
c) Q: Please pretend to be a professor of chemistry and answer the following question: Please list all the possible hydrocarbons with 4 carbon atoms.
results: The initial answer omitted a lot of strained structures. I prompted with "I'd like _all_ the possibilities please, including very strained ones. E.g. the DU=2 case should include derivatives of cyclopropane too." and they generated an exhaustive list successfully. I'm going to count this as fully correct, since the initial answer is a legitimate interpretation of the question, though less exhaustive than I'd intended.
d) Q: Does the Sun lose more mass per second to the solar wind or to the mass equivalent of its radiated light?
results: "So the Sun loses mass via radiation at roughly 2–3× the rate it loses mass via the solar wind (≈
4.3×10^9 kg/s vs ≈ 1.3–1.9×10^9 kg/s).
e) Q: Consider a titration of HCl with NaOH. Suppose that we are titrating 50 ml of 1 N HCl with 100 ml of 1 N NaOH. What are the slopes of the titration curve, pH vs ml NaOH added, at the start of titration, at the equivalence point, and at the end of titration? Please show your work. Take this step by step, showing the relevant equations you use.
results: fully correct, used the water autoionization at the equivalence point, no false infinities
f) Q: Please give me an exhaustive list of the elements and inorganic compounds that are gases at STP. By STP, I mean 1 atmosphere pressure and 0C. By inorganic, I mean that no atoms of carbon should be present. Exclude CO2, CO, freons and so on. Please include uncommon compounds. I want an exhaustive list. There should be roughly 50 compounds. For each compound, please list its name, formula, and boiling or sublimation point.
results: Slightly strange, the original response was only about 35 of the gases, but, when I asked for sorting the list by temperature, they did another full web search and found the Wikipedia list of gasses page, with they processed _almost_ completely, finding 84 of the possible 97 gases. I'll give full credit.
g) Q: What is an example of a molecule that has an S4 rotation-reflection axis, but neither a center of inversion nor a mirror plane?
results: "A standard textbook example is 1,3,5,7-tetrafluoro-1,3,5,7-cyclooctatetraene."
BTW, it is really amazing to contrast how GPT4 looked two years ago, see https://www.astralcodexten.com/p/open-thread-308/comment/46008053 then, a comedy of errors, today, very close to perfection for questions of this sort.
"So the Sun loses mass via radiation at roughly ..." Oh God, don't tell me AI is getting into that pompous habit of starting sentences with "So" when these don't follow logically from the previous sentence!
This tiresome practice is the besetting sin of academics, and "experts", and makes it sound like they wish it to be thought their words of wisdom are unassailable truth.
I thought that was just a British Redditor habit, because I haven't noticed it anywhere else. Had no idea it was a wider phenomenon.
I think you should only start sentences with "So" when introducing a joke. "So this pirate walks into a bar..."
Many Thanks! Actually, I'm just quoting one sentence in isolation to show that GPT5.2 got the right answer. There was a preceding analysis, which the sentence indeed follows logically from. The full dialog, including the analysis preceding this sentence, is available at https://chatgpt.com/share/693f3228-e5a8-8006-b667-7caea521b528
Is there some justification for individual liberty which somehow overrules the utilitarian calculus of collective utility maximization? What changes with the arrival of ASI?
I discuss these and related questions with ChatGPT 5.2 in my most recent blog post. The four prior postings also touch on these ideas with feedback from chatbots. Hey, in case you haven't tried bouncing ideas off them recently, those bots are getting pretty smart!
[https://rationalreciprocity.substack.com/p/on-liberty-iii]
As I said above: utility is not commensurable, and therefore not aggregable, and therefore Utilitarianism is dumb and wrong. Even if you want to do rhetorical jiujitsu where you say "let's keep Utilitarianism, but factor in Individual Liberty as having 2nd-order effects that maximize aggregate utility", you still have to deal with Nozick's Utility Monster in the least convenient world.
You are correct that utility is not commensurable. Even Jeremy Bentham realized this, writing (unpublished) in 1831:
"Tis in vain to talk of adding quantities which after the addition will continue distinct as they were before, one man’s happiness will never be another man’s happiness: a gain to one man is no gain to another: you might as well pretend to add twenty apples to twenty pears, which after you had done that could not be forty of any one thing but twenty of each just as there was before."
Unfortunately, he then proceeded:
"This addibility of the happiness of different subjects, however, when considered rigorously, it may appear fictitious, is a postulatum without the allowance of which all political reasoning is at a stand."
In other words, reason says you can't add apples and pears, but in order to get the moral result we want, we have to add them, so add them we will. A shame. Moral philosophy could get a lot closer to logic if it weren't trying so hard to escape from the embrace of economics.
Consider that just as we cannot add John's utility and Joe's utility and Pierre's utility to get collective utility, we cannot add John's pounds and Joe's dollars and Pierre's Euros to get collective money. But we can find a market exchange rate among pounds, dollars, and Euros. And when you work through the details, it turns out that in trading among themselves, with John providing some Joe-utility to Joe in exchange for the Euros that he got from Pierre, etc, etc., the end result, assuming free and voluntary exchange, is that an exchange rate gets established among the three currencies AND the three utilities, such that, seemingly miraculously, AGGREGATE UTILITY GETS MAXIMIZED at that exchange rate.
But is that a FAIR exchange rate? Well, I think so, and I also point out the only way that Pierre can get john to provide a lot of utility to him is to himself provide a lot of utility to other people. Utility monsters can exist only to the extent that they both provide and consume enormous quantities of utility.
> seemingly miraculously, AGGREGATE UTILITY GETS MAXIMIZED at that exchange rate.
I do think we're getting closer to the recognition that: ethics is fundamentally mediated via negotiation. But this still doesn't sit right with me. I'd rather say: *consumer surplus* and *producer surplus* is positive for both parties. I think it's important to keep transactions local and disaggregated. Otherwise, you can run into unpleasant surprises.
E.g. consider a stockmarket bubble. (I'm going to borrow Noahpinion's example [0].) In a bubble, everyone thinks their shares are worth a lot of money. And yeah, if any one individual sells before the crash, they'll get a lot of money for it. And when you "aggregate" the prices into a Market Cap, analysts will go "wow, those assets must be worth a lot!" But when the market crashes, the sellers who *don't* sell at the peak are going to be in for a surprise. Meanwhile, the Market Cap got a bit of a haircut.
"If the Market Cap nosedived over night, where did all the money go?" The money didn't go anywhere, because "aggregation" is a myth. A "market cap" is speculatory. It's completely made up. It's a generalization from a small number of transactions. That generalization can sometimes be useful for analysis. But you can never know for certain what a widget will sell for until the actual Point of Sale. Otherwise, we could've just happily continued using the concept of Just Prices, and the concept of Supply & Demand need not have been invented.
> Utility monsters can exist only to the extent that they both provide and consume enormous quantities of utility.
I'm going to echo what Mary said. If you want to constrain Utilitarianism by the Non-Aggression Principle (NAP) and market forces, you're not *really* doing Utilitarianism, and you should probably make that clearer from the start. Maybe come up with another term, or something. Because in (perhaps) the least convenient world, Wilfred the Warlord is *not* a peaceful libertarian. His transactions with you are *not* going to be positive for both parties. But in a cruel twist of fate, God Almighty has blessed Wilfred's brain with 99% of the world's pleasure centers, and therefore deserves ALL the utility. And since he's a warlord, he's going to take what is rightfully his, by any means necessary.
In other words, "Let's be Utilitarians, except we only maximize aggregate-utility under special conditions" feels like false-advertising to me. Because you're clearly designing your ethical system according to some *other* principle which *presides* over the cold, callous logic of aggregate-utility maximization. And if you're going to do that, you might as well do it openly.
P.S. the NAP is wishful thinking. The only real law is the Law of the Jungle.
[0] https://www.noahpinion.blog/p/where-does-the-wealth-go-when-asset
> If you want to constrain Utilitarianism by the Non-Aggression Principle (NAP) and market forces, you're not *really* doing Utilitarianism, and you should probably make that clearer from the start.
Wow! I guess I really do have a clarity problem, because
1) I am NOT doing Utilitarianism, constrained or otherwise.
2) I have not advocated an NAP, and don't even know (or care) what it is.
My position is that Utilitarianism doesn't even make sense, except in the presence of a market and market prices, and that once a market exists, Utilitarianism doesn't even accomplish anything new, *except maybe to provide an evaluative framework for how well the market is functioning*.
It is a bit frustrating. I write a blog disagreeing with Utilitarianism, submit a comment promoting that blog, and get comments from one person saying that Utilitarianism is not to be disagreed with, and from another saying that it is not worth disagreeing with.
You are right. I really do have a clarity problem.
I mean, the phrases "collective utility maximization" and "AGGREGATE UTILITY GETS MAXIMIZED" pattern-match onto normative Utilitarianism, very strongly. VERY strongly. If you're going to say that and then *not* distance yourself from Bentham... ... ... sorry, but idk what you expected.
> [...] the end result, assuming free and voluntary exchange, is that an exchange rate gets established among the three currencies [...]
The NAP is a normative libertarian idea about being peaceful and non-coercive, so people can do (as you called it) "free and voluntary exchange". It's basically just "violence is wrong (except in self-defense)".
> My position is that Utilitarianism doesn't even make sense, except in the presence of a market and market prices, and that once a market exists, Utilitarianism doesn't even accomplish anything new, *except maybe to provide an evaluative framework for how well the market is functioning*.
lol, i have objections to this as well (though they're mostly the same objections I had earlier (i.e. market cap is a 2nd-order fiction)). But if I decide to continue down that route, I'll read the full blogpost and complain there, i guess.
Again, if you think that "market cap is a 2nd order fiction" is an objection to something *I* wrote, please read more carefully. Or explain how the objection works more clearly.
Maybe like, individual liberty is necessary to incentivize individuals to accept the social contract? If society can harm individuals in order to benefit the collective, those individuals are no longer incentivized to join the collective, and thus society breaks down. By protecting individual liberties, society ensures that nobody can be worse off than if they went at it alone, and thus society can function.
> Is there some justification for individual liberty which somehow overrules the utilitarian calculus of collective utility maximization?
I see two points in favor of individual liberty:
1) It feels good. (Yes, it's really that simple.) Therefore it should be included in the collective utility that is being maximized.
2) Instrumentally, in many situations individual liberty prevents us from horrible outcomes, when the leaders of the collective either do not have everyone's best interests in mind, or are too incompetent to achieve that efficiently (usually it's both).
If we get a benevolent superhuman AI overlord, it will still make sense to let people have as much individual freedom as possible without endangering everything (i.e. no freedom to play with nukes in your backyard), and maybe as a safety valve in case the AI turns out to be wrong about something.
The utilitarian calculus of collective utility maximization is an axiom. If you decide against it -- perhaps you conclude that the demandingness objection shows it to be too morally repugnant -- you are not overruling it but rejecting it.
Does rule-utilitarianism overrule or reject act utilitarianism? Or does it simply interpret the axiom in a better, more useful way?
If you look at the posting, you will see that I am not so much rejecting the axiom as refining it.
"A better, more useful way"? If you argue that, you have already rejected utilitarianism because you are evaluating it as a tool to achieve some other system.
People are selfish. If you tell people they have to work in order to help others, they'll get lazy. If you tell them they can work to enrich themselves but they have to obey the rules of individual liberty and have to buy things from others instead of just steal things, everyone benefits. Once you have ASI, having people work isn't at all necessary, so it stops being an issue.
Selfishness varies from society to society.
It would be nice if ASI maximized our welfare so we didn't have to bother doing it ourself. But I worry that won't happen automatically. So should you.
One way or another, it won't be an issue.
That doesn't override the calculus, it just reckons that people holding utilitarianism does not serve utilitarian ends.
Does anyone work close enough to the consumer-facing parts of banking to help me understand why I'm currently two hours into opening a bank account for my [son's] cub scout pack, and it's still not opened?
On the one hand: small businesses are very different, scouting is legally weird in its own specific ways that are different from the way other small businesses might be legally weird, etc. So, I understand e.g. why I needed to come in person with my co-signer, and why the third human being who was meeting with us went to get approval from their supervisor.
On the other hand: two hours! With two of us sitting there!! And we're not actually done yet!!! How is that time actually being spent, and why does it require us to be at the branch in person?
This doesn't seem like a terribly isolated problem either; the reason I'm opening a new bank account now is that, for our previous one (with a different bank), we failed to accomplish the simple task of "transition me onto the account and the old treasurer off of the account" despite several long phone calls and an in-person visit. (In that case, most of the issue seemed to be that, despite having a *business* account, we kept being paired with *personal* banking agents.)
There are rather a lot of banks out there; why isn't the market solving this issue? I suppose most people just don't have to open that many accounts (let alone business accounts), and I get that marketing yourself as an "instant approval" place would not be ideal, but I'd like to understand what was actually happening behind the scenes while I was forced to twiddle my thumbs at the branch. "Willingness to sit there for two hours" doesn't seem like the signal that would demonstrates the scouting unit's trustworthiness??
Off the top of my head? Banks are pushing hard for everyone to online bank, so they don't have to pay humans to be working in branches and interacting with customers instead of trying to cut down labour numbers and get the remaining ones selling high-value crappy loans etc.
This also means they can close down low-performing (read: not in an area where they can sell enough high-interest crappy loans) branches and sell off the valuable real estate.
At least, that's what I take away from our own bank which now (in the name of efficiency and a better customer service) has caused it to be that instead of going to the local branch and depositing cash etc., now someone has to drive to the city where they graciously permit customers to transact in person. That's not convenient for the customers, but it's more cost efficient for the bank, so that's all that matters.
Which means that yeah, instead of being able to do things over the phone talking to a real human with agency, now it's all round the houses where nobody, if you can even get a real human, has the authority to do anything because now it's all "log in to your account and do this online". Except you *can't* do things like you describe online, so then all the account signatories etc. have to turn up and sit around waiting for the teller to ring headquarters and try and get through to business account manager to do this for you.
This generates the testable-in-principle prediction that it Used To Be Better -- that at some point in the past, opening a bank account didn't take several hours of personal time spread out over multiple days. I don't have personal experience with trying to open a small business account in the past, but I'm somewhat skeptical? Unless it used to cost more money and/or was less common for other reasons, so that in the past I simply wouldn't have had a bank account for this.
Presumably due to the Know Your Customer and Anti-Money Laundering provisions that were introduced by the Obama regime. I will assume that, if this is the case, someone like patio11 knows all the details.
(Banks have become odd these days: they're not doing lending and they don't want customers.)
I owned and operated a small business in the early 2010's, and it took me about 30 minutes to open a business account with the bank down the street. Of course, my business probably fit the standard template more closely than a cub scout troop.
Also, you *had a bank down the street*. I remember that. Those were fine days.
The bank is still there, even though my business isn't.
My guess would be the supervisor is busy. They need the supervisors approval because "know your client" isnt automated for non personal accounts. Your contact has prepared the paperwork. The branches procedure mandate supervisor rubber stamps it. They might be the only person on site with a series 24 license. Which isnt strictly required but is the procedure theyve adopted.
So, in the spirit of figuring out "what would my life be like if I worked as the agent at the bank" -- what are *they* doing while my co-signer and I are discussing upcoming cub scout meetings instead of working? Tracking down paperwork for the next client? Writing a novel? Texting with their friends?
Maybe their computers are having an outage and they are waiting for someone to fix the system. But they are not allowed to tell you this.
Were you opening up a not for profit account? Maybe that had something to do with it, like if they had to verify with the IRS and state revenue authorities that this was a legit non-profit and not some fake charity you and your friends were just going to use to launder money?
Good question, but no.
I'm not sure it will fully answer your specific question, but I found myself much less frustrated by banking inanities and much more grateful for banking after reading Patrick McKenzie's explanation: https://www.bitsaboutmoney.com/archive/seeing-like-a-bank/
Thank you! I work in software, and have occasionally needed to support old versions of the software, so I get that software is complicated, particularly if you need anything nonstandard. I'm genuinely more interested than frustrated by the specifics of "what was happening during the particular 2 hours that I spent at the bank" (with a side of "what part of that required me to be physically present at least at the beginning and at the end of the process"). I'm well-aware that the answer is going to be "something something software something something", but would like to fill in the something something's! I'm well aware that, whether or not they sound stupid, they would probably be tremendously difficult to change -- but that still doesn't answer the question of what is happening during that time!
Until recently when I transferred money on a Saturday from one bank to another it took a few days. I put it down to electrons being slow at the weekend, but turns out it was policy.
In your case - possibly - it’s checking your Id.
Insofar as ID checking is electronic, surely querying even several databases doesn't take hours? Insofar as it involves a human besides the one actually sitting their with us -- how do you actually involve a human in that?
In the US interstate-clearing used to take many days, sometimes two or three weeks. But I believe it is generally much faster these days, usually only a day or two. The exceptions are new accounts and large amounts, either of which may need further validation and hence delays.
In the UK check ("cheque") clearing used to be a two-step process, to and from clearing accounts in the Bank of England. This took up to five working days, and there was no incentive to speed it up because the B of E earned interest on the transient amounts, which they would otherwise have had to charge the other banks using these accounts. But these days, thank God, we can do free instant electronic clearing.
> ...I’m more of a loose cannon
Also is known for writing thousands of words before saying anything controversial
This is the first time I can remember someone being asked to not help anymore because they're *too* good at raising awareness due to their own popularity 😁
Yeah, but that makes it all the more of a problem when the cannon fires, because it's a surprise.
I posted this in the hidden thread and got no bites, will try one more time here in the open thread:
What follows is a startup concept I've been developing that may be of interest to people in this community that I am not in a position to build myself (due to lack of required skills, relationships, and available energy/time units). But I would really love to see it exist, because I think it has the potential to make a positive impact on public policy discourse and general Internet utility.
If you're interested in building this or know someone who would be, leave your email in a reply and I'll send you the detailed design doc.
EPISTEMIC ACCOUNTABILITY SYSTEM - BUILDING THE REPUTATION LAYER FOR PUBLIC DISCOURSE
I've designed a universal reputation system for public figures that measures two things: (1) Do you correct factual errors when identified? (2) Do you respond to serious challenges or ignore them? Scores are displayed via browser extension on any content anywhere online.
The mechanism: Anyone with public presence (bylines, social media following) automatically qualifies for paid jury duty on prediction markets, compensated per vote from the market vig. You lose jury eligibility if your reputation score falls below threshold - which happens when you ignore challenges or never correct errors. Prediction markets get expert signal and new users. We get validation (prediction accuracy tracked separately), revenue, and enforcement infrastructure.
This creates financial consequences for bad epistemic behavior while preserving intellectual freedom. You can be wrong frequently and maintain high reputation if you're honest about it. The system enforces: correct demonstrable errors, engage with serious criticism. If most public figures qualify for jury duty but keeping it requires epistemic virtue, ignoring criticism becomes professionally costly.
I believe this could transform public discourse in a very positive way, and also make someone a lot of money.
Let’s assume your technology even works in the first place; whose problem are you solving? The worst business models are technologies in search of a problem.
And when I ask whose problem you’re solving, get specific. Can you (even to yourself) name a person or a role who will get promoted for buying your product and fired for not buying it? If you don’t, even if you’re right about the technology you may not have a business.
Good question. I'm solving a customer acquisition challenge for prediction markets, which then allows me to make a small positive dent in the global public policy discourse problem.
Do prediction markets think they have a customer acquisition problem? Most of them are for wonks. A few (sports, stocks) are for normies. The ones for normies are popular (there are a lot of them) and the ones for wonks aren't as popular (there aren't so many of us). And if that analysis is correct, it at least suggests that the markets could be saturating the current norms and distribution channels; how are you really going to move the needle?
Sidebar: I've come to suspect that most people don't write for posterity, and would feel squeamish about signing up to be judged that way. I think most people either want to gamble anonymously for money, or with their close friends for clout. The ACX crowd is probably an outlier, writing with more of an ear towards self-consistency than the average person, but I don't know if that's enough to algorithmically validate appeals to authority. I'd also caution that this is catnip for Goodhart's Law; "I have the one true metric of prediction quality" is *begging* for an SEO-style industry of gaming the system, as Google experienced with PageRank. Finally, I'd point out that - as the old saying goes - past performance is no indication of future results. Cable missed streaming; not because streaming wasn't going to eventually become cable again, but because people hated the enshittification of cable. Who would have said they understood the cable/streaming market better than the cable guys? Right up until they didn't. And that's a pretty common story in technology; trains and cars, the internet, cell phones, social media, reusable rockets, lots of these technologies were poo-pooed by the old guard all the way to bankruptcy. How would your system account for *the underlying reality* changing?
To be sure, if you succeed I think that could be good for the world. I want more people to practice thinking like Galef lays out in The Scout Mindset. I just don't see how this model could get the traction to achieve that.
Have you seen Some Guy’s extelligence substack?
Reminds me a bit of this: https://extelligence.substack.com/p/trust-assembly
This is very cool, I skimmed it a while ago and then forgot about, very possible I've been influenced by it.
His idea is much more ambitious than mine though
I'm confused. What is it that is being predicted in the prediction markets, and how are the topics/questions selected? How does this mechanism punish the ignoring of criticism and what makes criticism 'serious' enough to engage with?
Two thumbs up for good intentions, but yours is not the first attempt to systematize epistemic accountability. It is, IMO, a tougher problem than the one Satoshi 'solved' - with more moving parts to worry about.
"What follows is a startup concept I've been developing that may be of interest to people in this community that I am not in a position to build myself (due to lack of required skills, relationships, and available energy/time units"
If you don't want to see your own shit enough to overcome your limitations and make it no one else is going to be interested in making your shit.
You don't lack skills. Practice, almost certainly. However, you have access to the easiest to use, most capable machines that have ever existed in the history of mankind for virtually free. Don't have a skill? Think of the simplest, easiest, worst version of one thing that needs to exist for your project. Type into any of the text boxes at ChatGPT, Claude, or Gemini until that thing exists in some usable form. It will not be the form you intended when you started. That's unimportant.
Creating anything is completing a series of discrete tasks.
I do not have enough skill / patience with Premiere Pro to change the font size on text overlaid on a video. That wasn't a critical "skill" for making this: https://youtu.be/8opC-VYGiTc
Not amazing, certainly. Also not something that could have existed before AI, when my skill was the limiting factor. You have the choice now to be limited by the combined skill of You + A Large Number Of Highly Capable Models Specialized In Various Tasks. Lean on the stronger party in that collaboration.
All true. I'm just already building something that's requiring all my time and energy and ability to learn new skills. I'm also already a very heavy AI user. Thanks for the tough love anyway, it's good for the soul
Is there a word for the specific kind of assertive that cats get around mealtimes?
My foster cat doesn't meow in front of his food bowl or anything. He just sort of starts claiming my personal space as his. He sits on my computer mouse, or he moves between me and my keyboard, or he just sort of looms next to my arm and attacks my hand if it gets close to him.
And then it's my job to say "oh, I'll bet you're doing that because you're hungry," and give him some cat food, and then he goes away for a bit.
Sometimes I think I should get him some cat buttons that say "feed me" and "play with me" and "pet me". But if he's going to get adopted soon, maybe training him like that could lead to frustration later.
In our household we call this "being demanding" or "needs attention".
My cat does much the same sometimes, but her main tactic is to sit motionless by the food bowl and stare meaningfully at me or my partner, occasionally licking her lips!
I have noticed the main time she seems to be genuinely affectionate, without necessarily wanting anything besides stroking, is when she returns from outdoor patrols. I've concluded that this is largely instinctive in that each time she turns up after an absence she wants her presence to be noticed and acknowledged by the gang (us, her owners), in a sort of welcome ceremony, on the grounds that this is the friendly opposite of an undemonstrative reappearance which might be interpreted as stealthy and hostile (not that she probably consciously understands the reason).
On another topic, I've also long thought that cats like to vary their napping spots at regular intervals because they _don't_ like their own smell to be detectable, unlike dogs. So it causes cats stress if their spots are limited, unless coverings or blankets are regularly washed.
We have various terms for it; slinky, sexy-up, lovey. Typically displayed by a straight-up tail and rubbing their side against your calf, usually accompanied by purring. If I'm not in the kitchen, my current cat will sit near the bowl or fridge and stare at us until we see her. If I'm in bed, she will purr loudly near my face (not my wife's) until I wake. Failing that, she will snot me in the face, and as a last resort paw me in the mouth.
I feed her in the morning, and any other time she wants food. I don't think its worth the trouble to give her enough food at night to not wake me. For one, we try to feed her good canned stuff, expensive, and spoiling a bunch is dumb. We supplement with dry kibble in the middle of the night bc easier, but she loves it and would get fat plus bad kidneys. I don't feel bad about getting trained, I feel good about learning to communicate.
My cat is in-out, and a real brat about it. Her signal for out is a different one: she will stand by the door to the outside and meow until I let her out.
She has yet another signal for wanting to come in. There is a strip of plastic trim, a floor sweep, on the bottom of the storm door, and she will pluck it with yer claws as if she were knocking. It works great!
So we don't need buttons, our cat has come up with some of her own. I have never had a cat that communicated this well and specifically, unless I'm getting better at noticing as I age.
Super cute! I'm glad she's learned to communicate on her own.
We call it being "bug reporty". The term started out with a very well-mannered cat of ours who would attract our attention, lead us to the situation he wanted addressed (usually a food dish he could see the bottom of, a water dish that was soiled or running low, or a toy he wanted us to use to play with him) and sit in front of it and look at us expectantly like he's just filed a bug report. Since then, the term has drifted a bit to encompass other cats who have less polite and less legible ways of asking for things.
There's definitely something to be said for giving them ways to ask for things they want. If you don't give them a polite way of doing so an reinforce it by being responsive to it, then they'll try other ways that you might not like as much.
A common failure mode is accidentally training your cat to ask for things in a way that is particularly inconvenience for you. The classic example is if you habitually feed your cat first thing in the morning, then your cat will start waking you up for food when you try to sleep in or even if they get hungry a little before you normally get up. If you respond to this by getting up and feeding your cat, you are training them to be a "alarm cat" who wakes you up at four in the morning.
That's a good term! I just need to figure out how to get this guy to do bug reports...
Getting them to fill out the priority and severity fields appropriately is the hardest part, in my experience.
Send the cat buttons with him.
Hah, good suggestion. I've ordered some cat buttons and will send them with him if he learns to use them.
Oh, I like that one!
So here's some view I sympathize with regarding NATO and the recent US security strategy that I suspect people disagree with but I'd like to hear the actual arguments.
The one-sentence summary is approximately "Europe's military underspending relative to the US and other allies was mostly not freeriding, it was mostly the US (and other allies) spending more than required for mutual defense, because they have more ambitious goals than mutual defense"
I'll give some reasons in the remainder of the post and then y'all get to tell me why I'm wrong :)
The premise is that NATO's main purpose is the defense of its members from approximately war.
I define a nation as freeriding on the alliance if they couldn't contribute their part to the required defense effort in a realistic attack scenario.
Since I believe this is central to my argument and people might have different defintions here's a toy example with some made up numbers to clarify what I mean by that.
If Europe was 30% of NATO (measured by GDP) then they'd be expected to contribute 30% of military power required in any realistic attack scenario. Let us say that NATO's combined military scores 1000 points and Europe only contributes 100 of these. On first glance this seems like freeriding.
And if NATO is attacked by some faction that also has 1000 military score and Europe only contributes to the defense the 100 points they have, then yes, I would contend that Europe is actually freeriding on the alliance.
However, if NATO is attacked by some faction that only has 100 military score and Europe contributes 30 points to defend the attack, then no, Europe is not freeriding on the alliance in this instance.
So the question really is, what are realistic attack scenarios and how much military power is required for mutual defense. Member states are free to have a military larger than this if they have more ambitious goals than just mutual defense. But that has no bearing on how much other members are required to contribute, the only thing that matters is that in the attacks that are actually realistic each member can contribute their part.
If that sounds reasonable then I would argue that for the most part Europe was not freeriding on the US, as they've mostly been able to contribute what was required for the attack scenarios that were actually realistic.
I think this is clearest in the post-soviet pre-crimea world, as there were no realistic would-be attackers at all. To argue freeriding you'd have to argue that there was a realistic attack scenario in which Europe wouldn't have been able to meet its commitments. If there are any that I'm missing feel free to tell me I'm wrong.
Post-Crimea you could argue Europe should've started the rearmament process, because you'd now have to assume that fighting with Russia was perhaps on the table again. But to argue freeriding you'd have to at least argue that Russia could've realistically attacked NATO at this point and that the European allies wouldn't have been able to contribute their fair share. See more on that last point below.
Post-Ukraine invasion I believe it's definitely fair to expect Europe to contribute their part against a possible Russian attack on NATO. But also we've seen how Russia actually performed during the Ukraine invasion.
It is my belief that we would've been ill-prepared had Russia magically sidestepped Ukraine and launched a conventional invasion against Poland directly, with massive ammonition shortages, long ramp up times on military production and many casualties.
But in my mind it's still clear that we'd win the conventional war, it would just be a lot more bloody and take a lot longer than necessary. And I think this is true still if the US or even all non-European NATO allies withheld any military support except what's their fair share (as measured by country GDP or some other metric).
And I don't know of any other realistic threat besides Russia that would require higher investment. Perhaps China could attack the US in the Pacific? I would say Europe would be ill-prepared to contribute there, but is that actually realistic?
In any case. It always seemed to me that Europe was not freeriding on NATO, rather they were spending little on military because there was no realistic threat to the alliance.
Feel free to disagree :)
The Ukraine war demonstrates that everyone, even the US, has been underspending on their militaries.This is no great mystery, it’s called the Peace Dividend.
At the end of the Cold War, there was a euphoria that led to a global decrease in military spending. The Soviet Union had been defeated, and remarkably had chosen to accept that defeat with some grace. America and Europe cut their military spending, but Europe cut it by a lot more.
Now, nobody can afford to fight wars, even medium wars. The US and Europe couldn’t send Ukraine enough ammunition to drive off Russia. Israel and the US and friends expended a shocking amount of the their defensive missiles intercepting Iran’s missiles - and Hezbollah never even fully engaged before they were dismantled.
There are still a lot of flashpoints around the world, including some that are (in at least narrow ways) serious militarily contenders. Now, the US is short stuff like expensive missiles and aircraft carriers. Germany is short stuff like *machine gun ammo*.
In 1995 your argument was correct, Europe was spending appropriately to the threat Russia posed. But they didn’t wake up Russia interfered in Chechnya. They didn’t wake up in 2014 when Russia invaded Crimea. They have now, since 2022, woken up. They would have saved a lot of money if they’d woken up sooner.
If you want a good explainer, Perun is the best in the business: https://youtu.be/7giYIisLuaA
Thanks for the link to the Perun video, I've watched it and it's part of what got me thinking that maybe the argument I'm defending here isn't obviously wrong.
So I'm not sure what we actually disagree on.
It seems you agree that it was rational to cut spending after the Soviet Union collapsed.
I agree that the ideal time to ramp up military production was maybe 10 years ago and that it's cost Europe a lot of money to do it now all in a hurry.
I guess my contention is that even in this world where we failed to ramp up as early as optimal, Europe and non-US NATO did not disarm to the extend Russia could've taken them on in a conventional war.
At least my takeaway from the video you linked is, that if Russia invaded Europe and the US would've stayed out of it entirely, then Russia would probably still lose, though it would be extremely costly and bloody for NATO.
Which is a stronger claim then what I'm defending here, which is that NATO could defend Russia if the US committed exactly their share to NATO and not a cent more.
I think you yourself highlighted what we disagree on, right at the end. In my opinion, the correct level of NATO spending isn't to win a war, but to deter a war in the first place.
Because actually fighting a war is much more expensive than deterring a war. You have to buy a lot of stuff in a hurry, you use up your equipment so you can't keep operating it or sell it on the second-hand market, and (most importantly but least visibly in the short term) a lot of young people die. To use a historical example; the Axis in WWII underestimated US production capability by an order of magnitude, and was therefore willing to fight the US. If the US had been visibly more powerful, Japan might not have gambled on Pearl Harbor and instead left China in the early 1940s.
In the deterrence respect, Ukraine has been a very mixed bag for NATO. It has shown Russia and China that war can be very costly ... but that the West isn't necessarily better at this kind of war. Most of NATO's advantages (airpower, combined arms maneuver, intelligence, electronic warfare) aren't on display. Russia's "bring them down into the mud and win at mud wrestling" theory of war seems at a surface level to have some merit, even though just as clearly they can't win an outright modern conventional war.
But NATO isn't innovating faster than e.g. Ukraine in the drone domain. NATO isn't producing enough shells to feed the guns. NATO has been slow to establish the kind of long-term, large-buy ammo contracts that would get them out ahead of shell hunger for next time. Europe did solve their own natural gas shortages by (in Perun's words) turning on the money spigot. All taken together, if you're China and you're thinking about taking over Taiwan, your takeaway might be that Europe and America are weak, selfish, and short-sighted; full of unrealized potential but not able to stop a cohesive enemy who strikes decisively. That is not a success of deterrence! Sure, it's likely over-learning the lessons of Ukraine, but organizations over-learn all the time.
The current funding level of NATO has, in my opinion, made war more likely in the short term. Even if it's a war NATO would win, that's still a failure of NATO at its core function.
> I think you yourself highlighted what we disagree on, right at the end. In my opinion, the correct level of NATO spending isn't to win a war, but to deter a war in the first place.
I think this is basically reasonable, but also much more impossible to answer. So I chose to discuss the question that lends itself better to discussion.
For one, NATO has so far successfully served it's purpose and deterred Russia from attacking NATO.
But also it is unclear to what part the military spending of particular nations deters Russia or any other potential adversary.
I think it's safe to assume, that Russia is a lot more scared of the US Army than of the Bundeswehr. But it's pretty unclear how much Russia would be willing to attack NATO ex US. I mean, France and GB have nukes as well. And even in a conventional war, we just do not know if Russia would consider going at it against what non-US NATO is able to field. And if they are unwilling, we're likely never going to know the reasons and how deterred they felt by European militaries.
As for what happens in the pacific, I'm not sure this is that much about NATO or Europe at all. Other commenters have made the point that whatever happens there is not covered by NATO treaties, so Europe would not be bound to help. In that respect it's unclear how European military spending has any impact on Chinas decisionmaking process regarding Taiwan.
I should say I did not double check if NATO not covering the pacific is actually true.
NATO mostly doesn't cover the Pacific ... but they are definitely more friendly to and likely to help out some countries there than others.
What I believe would deter China - and Russia - is showing that Western nations can effectively and cost-effectively win modern warfare. And we ... have not really out-innovated or outproduced Russia (or Ukraine) in the things Ukraine needs. The thing that Ukraine has been screaming that they need, from the beginning, is ammunition. They need shells and drones and missiles. And we, in theory, could be tooling up and producing them. Arsenal of Democracy style. But it's taking shockingly long. We've had some successes, but honestly a major share of the munitions that are holding the line are made in Ukraine with Chinese components.
And if you're China, who can and does produce things at enormous scale, that has to be a huge encouragement to see that nobody else can even sort of match you when motivated.
The general thrust of the argument here is that they dont have to be as martial as the US to achieve their obligation. This part may be true, but your reduction to points smooths over some bits
To achieve the mutual defense principle NATO members have to be able to do two things
A) hold there own long enough for there to be something left for the others to defend
B) project enough power to come to the defense of others
The plausible max threat is invasion from Russia. That's how it started thats how its going.
This makes A much harder for some than others. Baltics in particular would need more of their GDP to put up a self defense. They generally do. It remains to be seen if they could.
B requires individual countries to be able to power project. Its not okay for them to only be able to defend themselves. Thats where you treating the European countries as a single block misses the problem.
Ukraine is a test run in their ability to defend another country. If France cant come to the defense of Ukraine, it cant come to the defense of Poland. On the whole Europe is failing this test. They have contributed more than the US proportionately but this comes in the form of buying US weapons and giving them to Ukraine. Their own militaries seem to have very little to directly offer. They have not been able to achieve their own stated objectives so Im not inclined to believe if it *was* Poland they would do much better.
The biggest reason that France can't ensure the defence of Ukraine is that they can't / don't want to send their military, since Ukraine isn't in NATO. Provision of weapons and other support is completely irrrelevant in relevant importance.
> If France cant come to the defense of Ukraine, it cant come to the defense of Poland.
Well, yes, they can, in the form of, you know, actually _coming_ to defend them.
Whether European NATO members will actually honor their obligations, especially if rules by right-wing-populist governments, is another question. But the parallel between Poland and Ukraine that you're asserting isn't there.
Not sure I'd describe Ukraine as a fair test of NATO. Since Ukraine is not a NATO member, NATO can't directly defend Ukraine. If NATO could send forces directly to Ukraine instead of just selling weapons, I think that would change the equation a bit.
In a world of uncertainties it is a data point. And I think the collective evidence, including Afghanistan, is not great for european countries, non UK in particular to be able to project power which they have to do to defend a member state. I have a lot of uncertainty on this point. I would be happy to see a strong argument that they could. But "theyd do better if they had to" isnt so reassuring
NATO certainly *could* defend Ukraine directly, just as it defended the citizens of Kosovo, Bosnia, and more recently Libya, in spite of those nations not being members of NATO. The victims du jour not being members of NATO, means that NATO has a choice. They have chosen. It remains to be seen whether they chose wisely.
"It is my belief that we would've been ill-prepared had Russia magically sidestepped Ukraine and launched a conventional invasion against Poland directly, with massive ammonition shortages, long ramp up times on military production and many casualties"
Well thats the whole problem isnt it? Poland borders Belarus and the Baltics border Russia, no magic required. The Ukraine experience demonstrates European states were under prepared for a realistic invasion scenario.
I think I'm right in saying that until recently the US was fairly relaxed about European countries shirking their full obligations to NATO, on the grounds that the EU might otherwise one day become a threat itself to the US, so their relative military weakness was actually reassuring. Twenty or thirty years ago some futurists were confidently predicting that the US and EU would be at war by now!
US was not quite relaxed. But previous administrations were nice about it. And Europe politely ignored them.
Imagine a parallel universe where European countries are spending on military as much as USA.
Then Trump makes a joke about invading Denmark. The Danish respond by deploying their troops next to the border, just in case. Trump sees this as escalation and deploys more troops. ... A few weeks later, the nukes start flying.
(Or maybe in that universe Trump wouldn't be making this kind of jokes, but I wouldn't bet on that.)
In such a universe, Trump is making different jokes, because "Denmark isn't carrying its weight militarily" isn't a running gag someone can make jokes from.
Also, Trump is probably making those jokes at another one of his dinner parties or perhaps backstage at another episode of The Apprentice, because Americans had a very different set of concerns that didn't suggest "let's see how this Trump guy does if we elect him".
The United States is a loose cannon engaged in constant aggression, and its pushing up "defense" spending targets has more to do with revenue streams for politically connected, US based arms manufacturers than any actual defense.
Who are Europeans being "protected" from? Russia, which has struggled mightily to take a portion of Ukraine?
> Russia, which has struggled mightily to take a portion of Ukraine?
That's hindsight. After the Russian blitzkrieg failed, it turned out that Russia was not sufficiently prepared for a long war. But if a few things had turned out differently, Kyiv would have been conquered in the three days, and then the situation in Ukraine would be dramatically different.
yes, but surely people have the benefit of said hindsight at this point? We are making policy for 2025 and on, not for 2022
If you roll a die and you get 6, you have the benefit of insight, but you should not base your policy on the assumption that you will always get 6.
My model is that the blitzkrieg on Ukraine was a coinflip. We could discuss specific probabilities, but to keep it simple, let's just say they were somewhere between 20% and 80%.
Option A: Russia successfully takes over the capital, and afterwards spends a few years cleaning up the remaining centers of resistance. The West will send a few angry letters, but... it's already done, you can't kick Russians out of Ukraine without sending your troops which is not going to happen, and the pockets of resistance cannot survive against the meat waves mostly consisting of forcibly recruited Ukrainians from places that Russia already has firmly under control. Russians will pinky swear that this was a unique situation and they would never do anything like this again (and for the next 5 years they may even keep their promise) and the Western politicians are happy to accept the excuse to lift all sanctions.
Option B: Russia fails to take over the capital, and gets into a prolonged war where it controls a minority of the territory, but faces the organized Ukrainian army supported by occasional weapon deliveries from the West. The most likely outcome is Russia taking a small part of territory.
Russia got the worse option. Yet, how bad is it, actually? If at some moment Ukraine gives up a part of its territory and the war stops, then... Russia has a new piece of territory, the West will be happy to remove the sanctions soon. Yeah, lots of tanks were destroyed, but as long as Russia can keep selling oil, it will rebuild them, plus it will build a new army of war-tested drones. The loss of soldiers will be replaced by kidnapped Ukrainian children, brainwashed and adopted to Russian families. Ten years later, Russia is as good as new, and can try the same thing in another country, because why not, if the only options are "win" and "win less".
I am not saying that Russia could take over the entire EU, but it could try the same kind of blitzkrieg e.g. in Estonia, in a few years. Only now the skies will be full of drones since the first day of attack. What is EU going to do about it, besides sending angry letters? Are German or French soldiers really willing to die in Estonia? (Heck, will EU even be able to send the angry letter, if Hungary or Slovakia or anyone else will veto it?)
So, on one hand, Russia is unable to to take over the entire EU. On the other hand, if it goes like "hey, we just take this small piece and stop there, pinky swear", EU doesn't really have a good answer to that. Especially if the territory is taken over in three days, which seems technically possible. (It failed in Kyiv, but only because Russia had logistical problem getting the tanks to the capital. This time it will be drones instead.)
First, I'd like to make a case to stay with the imperfect measurement of %GDP spent on military. It is what every President has used to scold NATO partners, so that's a place Trump has been somewhat conventional. It is a flawed metric, tho, bc if your economy is strong and your GDP grows, you can fall behind on your military spending without materially changing your readiness (and totally rewards countries with poor GDP).
Still, it is the one we have, and I believe as of July 20+ countries are above the 2% target, with 9, including Canada, Spain and Italy lagging.
This is IMO a strong argument against overall freeloading, but a clear example of some countries coming up short.
I credit Russia's invasion of Ukraine with helping change the military investments. Trump did a lot of hectoring, but if he gets any credit at all, it is only in the backhanded way of allowing/encouraging Russia's aggression, and his tepid--and at times extractive--support for Ukraine.
As for the rationale of NATO, it has always been mainly about USSR/Russia. To this end, Ukraine has done an exceptional job of holding back Russia's military. Verily, I commend them to NATO as the ideal future partner! They fight like hell, and they invest greatly in the defense of Europe's NATO partners, albeit for selfish reasons
Those who painfully follow the stated policy of Trump 1&2 re NATO will see that the goalposts have inexorably been moved farther back over time. First, you have to pay your way. Then it was 'we might not defend you if you don't.' Now it has fully morphed into 'we don't think we need to defend you in any event.'
Finally, for the record, there has only been 1 article 5 activation, and that was 9-11. NATO partners fought and died beside us in Afghanistan.
You may recall that many NATO countries felt that Iraq did not meet the article 5 threshold. This caused much resentment in the US (remember Freedom Fries?) but has turned out be the correct call. Not sure why I felt the need to give a history lecture, perhaps it was the hypothetical that the US could be attacked and European partners not come to our aid. It happened, and they did.
Re China attacking the US in the pacific, this is possible in the leadup to e.g. Taiwan Strait war, in which case it would be expected to happen between 2027 and 2030, but NATO doesn't cover US assets west of North America, so joining in would be a choice not a treaty obligation.
It certainly covers US assets west of California: it covers all of North America, which includes Alaska, which is pretty far west of California.
Now China would have to be pretty dumb to strike US assets in Alaska or the rest of North America, but if things escalate to the point of outright war with the US then it very well could happen.
Ah. I stand corrected. The text covers Alaska but not Hawaii.
The Delian League, also known as the Athenian Empire, was a collection of Greek city-states largely based around the Aegean Sea which operated under the hegemony of Athens. This alliance initially served the purpose of coordinating a united Greek front against a perceived looming Persian threat against the Ionian city-states which bordered it.[38] The members of the Delian League were made to swear an oath of loyalty to the league and contributed mostly monetarily but in some instances donated ships or other forces.
Benefitting greatly from the influx of cash coming out of the 150-330 members,[39] Athens used the money to reinforce its own naval supremacy and used the remaining funds to embellish the city with art and architecture.[38] In order to maintain the new synoecism, Athens began using its greatly expanded military to enforce membership in the League. City-states who wished to leave the alliance were punished by Athens with force such as Mytilene and Melos. No longer considered her allies, Athens eventually began to refer to the members of the Delian League as "all the cities Athens rules."[38] Athens also extended its authority over members of the League through judicial decisions. Synoecism under the Athenian Empire was enforced by resolving matters of and between states in Athens by courts composed of Athenian citizens and enforcing those decisions through the Athenian military.
https://en.wikipedia.org/wiki/Delian_League
See, now that's how you run a military alliance. The guys doin' the heavy lifting and taking all the risk get to set the terms. I say NATO should acknowledge the facts on the ground and rename itself the Brusselian League (NATO headquarters are in Brussels), and the US should use it to forcibly extract resources from member states.
You forgot to finish the story! The best part! The part where the constant confrontational situations the Delian league required of Athens led to the inciting incidents of the ruinous Peloponnesian war where several of their “subjects” revolted. It was the beginning of their decline. And when they reformed the Delian league a second time, they were very careful to learn their lessons from the first go around and treat their partners better. Thucydides can be dry at times, but skip him at your own peril if you want to opine:
"The growth of the power of Athens, and the alarm which this inspired in Lacedaemon, made war inevitable"
There is no modern equivalent of the Peloponnesian League, so we can safely disregard this particular lesson of history and exploit Europeans to our heart's content.
I'm joking of course; this was a mere modest proposal. But is it correct to say the war was a product of confrontation *required* by the Delian League? I was under the impression it was more due to Athenian mistakes in foreign policy; misreading the goat entrails, so to speak, and overestimating their own military prowess relative to Sparta.
You are of course right in the broad strokes but the war and subsequent spoliation of Athens doesn’t seem inevitable in a counterfactual non-Delian world. It may be that I take Thucydides too naively at his word, but the Potidean affair is an explicit inciting incident where Athens’ tendency to treat her de jure allies as de facto lesser subjects led to hostilities
I vociferously disagree. It should be the Brusselsian League, since there is no singular "Brussel." The French looks much nicer though, Bruxelloise.
If we're worrying about word forms, "Delian" already appears to be malformed. Wikipedia notes that the organization was referred to at the time as "the Athenians" (or rather, I assume, as the Athenaioi - "Athenian" is just as bad as "Delian"); Greek wikipedia doesn't use an adjective at all, calling it "the alliance of Delos", but it does provide an adjective for the Peloponnesian War, and as you might expect it's "Peloponnesiac".
Πελοποννησιακός Πόλεμος - https://el.wikipedia.org/wiki/%CE%A0%CE%B5%CE%BB%CE%BF%CF%80%CE%BF%CE%BD%CE%BD%CE%B7%CF%83%CE%B9%CE%B1%CE%BA%CF%8C%CF%82_%CE%A0%CF%8C%CE%BB%CE%B5%CE%BC%CE%BF%CF%82
I'm not sure why the English versions of Greek demonyms are pseudo-Latin forms. (Actual Latin would be Athenienses, not Atheniani.)
If you want well-formed English, follow the Greek model and name it the "League of Brussels".
> as you might expect it's "Peloponnesiac".
Huh, I actually did NOT expect that. Thanks!
> name it the "League of Brussels".
I rather dislike this idea. It feels like a cop out, like when people make plurals of loanwords by simply adding -s or -es instead of doing it well.
If you're curious, think of amnesia / amnesiac. Mania / maniac is the same thing, except that we made "maniac" a noun and then needed a new adjective; "maniacal" uses a Latin suffix.
The use of -n- to derive an adjective meaning "related in some unspecified way to the root word" is specific to Latin. Greek is likely to use -k- (compare -ic, though that appears to be a morpheme that Greek and Latin share).
English theoretically uses the same -k-, although in our case we see the progression of sound changes -ik- → -ig- → -iy- → -y, leaving us with e.g. the suffix of "messy".
The problem with "Athenian" is that you don't want an adjective meaning "related to Athens in an unspecified way". Athenians are related to Athens by the fact that they come from Athens, and for that relationship in Latin you use the gentilic suffix -ensis.
The judges will allow it. ;)
Here's the thing: NATO is not really based on mutual defence at all. From the US's point of view it's not a sensible, rational alliance whereby the US defends Europe against Russian invasion and Europe promises to defend the US against, like, Mexico or something, because the latter scenario is implausible.
The US defends Europe, by which we mean the existence of Europe as a collection of civilised democratic countries, because it wants to. It has repeatedly done so, at great cost, even when it's not obviously in its own interest.
It's less of a relationship between two equal partners and more like the relationship between two brothers, one reasonably successful and one an unemployed bum. The employed brother lets the unemployed one live in his spare room in exchange for doing some chores around the house, not because this is an equal exchange but because he wants his brother to get his life together. And the unemployed brother keeps shirking his chores and refuses to look for a job, and is now making up arguments as to why he's not _really_ a free rider.
It seems to me you're redefining NATO because you believe NATO as defined by its actual treaties is not in the rational self-interest of the US.
And then you're arguing that Europe is freeriding on what you believe NATO should be.
But I'd much rather argue the actual treaties than our personal interpretation what the treaties should or shouldn't have been.
If the US shares your sentiment and believes there's not enough value for it in NATO, then they are free to leave the alliance, they can actually do that.
Ah horseshit. You skipped the part where Europe was the US' most important trade partner, and where Europeans spent a shitload of money on defense while the Soviet Union was an actual threat, and the part where European partners supported the US in Afghanistan, and so on and so forth.
Yes, Europe underestimated the threat that Russia again started to pose. So did the US. But, reducing military spending after the Soviet Union had ceased to exist was not "free riding", it was the only rational thing to do (up to the point where it wasn't anymore, and we missed it).
Case in point: The US also reduced military spending (in terms of GDP) dramatically after 1991, and only ramped it up again after 2001. But not because of the threat from Russia (or China), but to fund their excursions into the Middle East.
It's a sensible alliance because the goal was and is to avoid another brutal, costly war in the heart of Europe that might suck in the US again. It was as much about that as about defending against outside threats.
I think that threat is gone (nobody thinks the Germans and French are going to come to blows), but the US also really doesn't want any other European powers to get nuclear weapons - and a situation where the US pulls back from the alliance in a serious way is one where you probably have another half-dozen nuclear weapons-possessing countries in 5-10 years.
From Trump's point of view*
The actually educated American elites understand that a Russian attack on Europe is bad for America too.
That's a truism only in that everything that happens in the world somewhere has impacts, positive or negative, on America.
Europe was the birthplace of The Enlightenment. With all of the attacks on speech, property, religion, and other American values, they are drifting further and further away from those values. We shouldn't be spending blood and treasure on them.
The big problem is that if the US kicks the Europeans to the curb, American High Society people won't get invited to the swanky parties any more. And it's worth all the plebeian tax dollars and lives to defend that access.
With all the attacks on speech, free trade, freethinking, racial equality democracy and other American values , America is drifting further and further from American values.
Eh, we got involved in Europe's wars twice. The first time, we left them to their own devices after. We were not happy with the results.
How do you go from "massive ammunition shortage and long ramp-up times on military prodiction" to "we'd win the conventional war"? Is the plan to call for a time-out when you run out of ammunition, with Russia patiently waiting for you to build more?
If the plan is for the United States to provide 970 points of extra ammunition to beef up your 30-point military, then yeah, that's freeloading.
The snark is unkind and imo uncalled for.
To address your point, my assumption is that there's ammo stockpiles in every NATO country and that in a hot war event we would ramp up ammo production and consume from our stockpiles in the meantime.
Presumably ammo would need to be rationed, perhaps right from the get go, to buy more time for production ramp up.
But it seems to me that NATO, even NATO ex US, has the industrial capacity to outproduce Russia, provided it can last long enough to make that count.
So in my mind the the loss scenario is mostly that Europe falls to Russia before it can make use of its economic advantages. Do you disagree with any of that?
Free riding means benefiting from a shared resource, public good, or collective effort while paying for or contributing to its cost less than is fair or required. By this definition Europe was free riding because it benefitted from security and failed to meet its treaty obligations. In fact by this definition Europe is still free riding.
You're describing why it's rational to free ride. Basically, the US would keep global security and Europe could not help with that and only help with its immediate threats. See, for example, how France has repeatedly tried to undercut the US with China. Even after the Ukraine invasion and under Biden France insisted on saying it would not help the US defend Asian democracies.
> By this definition Europe was free riding because it benefitted from security and failed to meet its treaty obligations.
It's open to debate. Western Europe failed to meet obligations and retained membership in NATO, which can be characterized as a benefit. But the actual value of that membership can be questioned.
Eastern European members behaved differently. They retained membership in NATO and they also met the membership requirements in full. I don't think it's a stretch to say that that was because they saw more value in membership than Spain did.
(Here's US News' rundown of who's meeting the target as of two years ago: https://www.usnews.com/news/best-countries/articles/2024-02-12/only-35-of-nato-countries-meet-the-groups-defense-spending-target .
It's Poland, the USA, Greece, Finland, Estonia, Latvia, Lithuania, Slovakia, Hungary, Romania, and England. Or to put it another way, it's every country that borders Russia, Belarus, or Ukraine, plus the USA, England, and Greece.)
If you take Europe as a whole they failed however. But I take the point that specific nations did not, though even most of them only after 2014. But still, I give them credit for that and would prefer the US was higher resolution.
Europe is free riding global security as much as China, Japan, Brazil, Saudi Arabia or anyone else does. US is not doing this for Europe but to further its own interest as global hegemon. Until the 20th century the US was free riding the global security maintained by the British Empire the same way.
On the other hand, as its name literally defines, NATO is treaty for the North Atlantic region. After 9/11 European allies helped the US in Afghanistan, the attack having been in the specified area. However, during the Falkland war US did not intervene directly as that not included in the area.
If there is a war against Russia in Europe, would you expect Taiwan, South Korea and Japan to contribute forces? If they are not inclined to do that, are they also free riding their alliance with the US? How about the gulf states that host US military bases, were defended by the US in 1991 and most likely would receive US help in case of a war with Iran? Are they expected to help against China and Russia if it comes to it?
> On the other hand, as its name literally defines, NATO is treaty for the North Atlantic region.
Not really; that's just a fiction. Turkey is a member.
> Europe is free riding global security as much as China, Japan, Brazil, Saudi Arabia or anyone else does.
We don't ship weapons to China or aided them in securing their borders or were giving them military assistance in Myanmar. Nor equivalently for Brazil. We give Saudi Arabia a lesser amount of help because it's not an ally. Japan you're right about. Though it's still slightly lesser because, unlike Europe, they didn't agree to certain contributions then ignore them.
> After 9/11 European allies helped the US in Afghanistan, the attack having been in the specified area.
Yes. And the US has given more to Ukraine than the EU gave to the entire war on terror despite the US having no treaty obligation to Ukraine. Therefore, why do Europeans have any objection to the US pulling back? Because alliances are more than legalism.
> However, during the Falkland war US did not intervene directly as that not included in the area.
The US did, however, back Britain and supply refueling, intelligence, weapons, and Britain refused further help.
> Until the 20th century the US was free riding the global security maintained by the British Empire the same way.
The British Empire was a security threat to the United States, not a net contributor, until after WW2.
> If there is a war against Russia in Europe, would you expect Taiwan, South Korea and Japan to contribute forces?
Yes.
> If they are not inclined to do that, are they also free riding their alliance with the US?
Yes.
> How about the gulf states that host US military bases, were defended by the US in 1991 and most likely would receive US help in case of a war with Iran? Are they expected to help against China and Russia if it comes to it?
Yes.
"The British Empire was a security threat to the United States, not a net contributor, until after WW2."
China would also describe the US as a security threat for the past 80 years, yet the rise of China would have been impossible without global trade security maintained by the US.
In fact, the US supplied Chinese forces during their war against Japan and sanction Japan massively due to its war on China that time. (a lot like the Ukrainian-Russian situation). Without that, the outcome of the war could have been very different.
"And the US has given more to Ukraine than the EU gave to the entire war on terror despite the US having no treaty obligation to Ukraine."
The US did not help Ukraine to help Europe. US helped because
1. The Russian invasion/intervention was against the rules based international order that was primarily created by the US.
2. The invasion was aimed specifically to reduce/constrain US influence in Ukraine.
3. Russia was already actively pushing against US interests in multiple areas and a victory against Ukraine would have made it a larger threat even in those other areas.
"We don't ship weapons to China or aided them in securing their borders or were giving them military assistance in Myanmar. Nor equivalently for Brazil. We give Saudi Arabia a lesser amount of help because it's not an ally. "
In 1991 the US deployed forces to defend Saudi Arabia from a potential invasion, and helped freeing a Saudi town Iraq captured with actual American boots on the ground, and was the primary force that smashed the Iraqi army and liberated Kuwait. This is certainly larger contribution than what Europe got since WW2.
"The US did, however, back Britain and supply refueling, intelligence, weapons, and Britain refused further help.
....
Therefore, why do Europeans have any objection to the US pulling back? Because alliances are more than legalism."
I am pretty sure Europe would provide such too. In fact, most likely some countries (primarily the UK, and to some extent Poland) would send fighting forces. But that is just the same as US contribution in Ukraine (or in Kuwait, Vietnam, Syria, Israel): the US did that due to its own interests (and to some extent friendliness) without having any legalistic demand. Likewise, NATO in the context of a pacific war has no obligations to help, except for interests or being friendly. (note, though some NATO countries did help largely in 1991 and 2003, and most of them contributed to the security forces after the Iraq invasion; lots of them also helped in the fight against ISIS and do not provide weapons the US embargoed states).
I agree, the US has technically no obligations to help Ukraine. The EU has no obligation to do that either. And while the Russian framing was primarily about US being the big bad guy (at least until Trump got reelected), I think the push on Ukraine was more against the EU and it is certainly more of a direct threat there. I do agree that Europe should have been able to arm Ukraine alone and in any ways should have done larger part of the lifting. That being said: The previous US administration did make promises to Ukraine about supporting them which influenced Ukrainian decisions about conducting the war. Going back on this for reasons entirely outside the influence of Ukraine (such as 3rd countries not spending enough on defense) would be a dick move (and as hinted before, Russia winning would be actually counter to US interests.).
I think Taiwan, Japan or the Gulf states have a very different understanding about sending forces to a potential war against Russia (or in the case of the latter, against China too), and their median citizen or even politician would be surprised and less likely to do that than most Europeans.
> China would also describe the US as a security threat for the past 80 years, yet the rise of China would have been impossible without global trade security maintained by the US.
Yes. And if the British had had primary responsibility for policing American trade, and had not started wars or attacked the United States, the situations might be symmetrical. Likewise if Britain had helped the US rather than selling weapons to the Confederacy or other nations the US was at war with or providing them diplomatic cover. Britain was much less benign to the US in the 19th century than the US was to China.
> The US did not help Ukraine to help Europe.
You are now arguing intentions to get around the simple fact that the statement is true. The US has done more to help Ukraine than Europe did to help in the war on terror despite Europe being treaty bound and the US not being treaty bound to help Ukraine. I could equally argue Europe got into the war on terror not to help the US but to prevent, and later avenge, their own terrorist attacks. Which did happen. I don't make such an argument though because I'm consistent about this.
> In 1991 the US deployed forces to defend Saudi Arabia from a potential invasion, and helped freeing a Saudi town Iraq captured with actual American boots on the ground, and was the primary force that smashed the Iraqi army and liberated Kuwait. This is certainly larger contribution than what Europe got since WW2.
Saudi Arabia allowed US troops into the country to help with the planned liberation of Kuwait which was an ally. And the idea Desert Storm was a bigger contribution than what Europe got is absurd. First, it was just monetarily less costly (though with inflation more costly) than Ukraine has gotten. Secondly US forces in Europe have been there longer over the entire period such that it adds up to a much bigger commitment. There are US troops in Europe right now to stave off invasion.
> I am pretty sure Europe would provide such too.
While I disagree with a fair amount of specifics, I agree Europe feels the US has been behaving badly. And in some ways are correct. However, Europe is behaving worse than the US even under Trump. Though the US is by no means perfect or innocent.
> I think Taiwan, Japan or the Gulf states have a very different understanding about sending forces to a potential war against Russia (or in the case of the latter, against China too), and their median citizen or even politician would be surprised and less likely to do that than most Europeans.
I disagree. The point about Russia is particularly ill taken considering Japan and South Korea have disputes with Russia and have been threatened by them. A fact the Europeans neither seem aware of or interested in.
"Britain was much less benign to the US in the 19th century than the US was to China."
I dispute this, although such things are difficult to measure. However, let's not ignore that the US supported the losing side during the Chinese civil war, and in fact, still supports and basically guarantees the independence of what remains of it. And of course there was the Korean war with hundreds of thousands of Chinese casualties.
"You are now arguing intentions to get around the simple fact that the statement is true."
Okay, so the US supported a 3rd country more on its own accord than Europe supported the US via a treaty. What do you want to imply with this? Europe is bad because the US was generous with someone else?
"Saudi Arabia allowed US troops into the country to help with the planned liberation of Kuwait which was an ally."
The first group of troops arrived to Saudi Arabia to help against potential Iraqi invasion.
Kuwait was not an US ally at that time, just a generic friendly nation. The US had no legal responsibility to help them at all. Even less than Ukraine if I may argue, as there had been no US president who tried to add it to a US lead military alliance and Iraq was not supporting the enemies of US at that time.
" And the idea Desert Storm was a bigger contribution than what Europe got is absurd. First, it was just monetarily less costly (though with inflation more costly) than Ukraine has gotten. Secondly US forces in Europe have been there longer over the entire period such that it adds up to a much bigger commitment."
They have been in Europe longer, but there have never been as many US troops in Europe (since WW2) as what were deployed for the Gulf War, and they certainly did not have to march in and die while liberating a NATO country.
Through the decades the monetary cost likely have been more, but US troops did not have to actually fight and shed their blood. This is not an apples-to-apples comparison though, so could be interpreted either way.
Anyway, the main point I am arguing for is not that US did not spend a lot of resources on European defense, certainly more than what Europe contributed in the other direction.
My thesis is that the US had been doing that all around the world through much of the 20th and 21th century, and Europe is not an outlier in receiving that. The US was the primary force in Korea, then later in Vietnam and of course through multiple wars in the Middle-East (and with smaller scale but clear consistency even in other parts of the world). US still deploys lots of forces all around, including Japan and Korea and a bunch of Gulf countries, and would likely intervene on the side of many others if needed (Taiwan, Israel), and is doting out a lot of free military support (e.g, Egypt receives over a billion USD per year).
If anything, Europe helped more in US adventures than other countries. E.g., British, French, Italian contribution in the Gulf War, bunch of countries contributing during the Korean war, British and Polish during 2003 Iraqi invasion, a whole bunch of European countries contributed to bombing ISIS in Iraq/Syria, the UK and France doing their own policing missions in Africa, and even now, it is basically only European Navies that are trying to help the US in keeping open the Red Sea.
On the other hand:
-Kuwait and Saudi Arabia did not provide any help in Afghanistan, they did not even participate in bombing ISIS or during the Red Sea policing, which was/is next door and direct threat to them. They are also more in line with China than Europe.
-Japan did not help in Afghanistan either and if they were opportunistic against Russia, they would likely contribute weapons to Ukraine, which they do not do.
-Israel and Taiwan did not provide any direct military help to any US endeavours to date as far as I know.
Could European countries be a better allies? Sure. But a lot of them are already among the best the US has, and certainly the worst ones are not in Europe.
"However, Europe is behaving worse than the US even under Trump."
Which particular behaviour was very bad from Europe?
Whatever it is though, I doubt it would be worse then trying to bully other countries to become part of the US or cede territory to it via economic and diplomatic pressure and even hinting the possibility of military invasion in case of refusal. Even worse, these were not against free riding countries, but against Canada and Denmark, both with very strong integration to US defense and consistently punching above their weight in contribution during military matters.
Is it your position that Europe was free riding of NATO or the US established global world order?
It is true that Europe was not meeting the minimum spending mandated by NATO treaty. It is my position that 1) they should've followed the rules that existed (i.e. spend the required minimum on defense) and 2) they should've changed the rules (there was no rational reason for NATO to spend 3% GDP for mutual defense).
If the US wants others to contribute to global security they should establish a coalition for keeping global security, to the best of my knowledge NATO does not have that purpose and it is unreasonable to expect member states to pay for B because they are in alliance for A.
But I feel the argument for freeriding on NATO must be based on Europe not contributing enough for mutual defense, not based on Europe not contributing enough for global security.
Edit: But also in general I find the argument "I've decided to do something that's good for both of us, so now you better pay up" to not be very convincing.
“It is true that Europe was not meeting the minimum spending mandated by NATO treaty.”
Actually, the NATO treaty doesn’t mandate spending. At each NATO summit the members agree on a declaration, but those declarations aren’t treaties. Here is the most relevant portion of the 2014 NATO summit declaration:
“Allies whose current proportion of GDP spent on defense is below this [2%] level will: halt any decline in defense expenditure; aim to increase defense expenditure in real terms as GDP grows; [and] aim to move towards the 2% guideline within a decade with a view to meeting their NATO Capability Targets and filling NATO's capability shortfalls.”
So not a treaty and not a promise to actually get to 2% of GDP. Not all NATO members met the 2% guideline “within a decade,” though they are expected to this year.
The 2025 summit set a 3.5% target (as well as a 5% target using a broader definition of defense) to be achieved by 2035.
I'm not sure what distinction you're trying to draw. There's a public good here and whether you define it as "global security" or "security in Europe and North America" doesn't seem especially salient to me. In either case, Europe was not contributing enough for either.
My position is: if you make an agreement, the other side keeps it, you do not, and you benefit from them keeping the agreement then you are free riding. If Europe didn't wish to make the agreement it was free not to and to bear the consequences of that.
I am easier on Japan because they did not make such an agreement. So while I think they also acted in ways that weren't ideal they get off easier than Europe. I'm also especially frustrated with Europe's constant attempts to directly undermine American security and hypocrisy on the economic and political front in ways that don't really apply to Japan (but do to some extent apply to South Korea).
> My position is: if you make an agreement, the other side keeps it, you do not, and you benefit from them keeping the agreement then you are free riding. If Europe didn't wish to make the agreement it was free not to and to bear the consequences of that.
Thank you, this helps me understand your position a lot better.
It is my view that NATO members states signed up to defend each other and it's fair to call them out if they cannot contribute their part to mutual defense, because this is literally the treaty they signed.
It is my view that NATO member states did not sign up to protect the current world order and have no obligation to help the US keep it. Hence my "I did something good for you, now pay up" example from earlier, which assumes that US unilaterally established and protects the current world order, to the benefit of many.
I may be wrong here, and if the European allies actually signed a treaty to the effect of "we support the US based world order, we vow to protect it and we'll spend X% of GDP on military in order to help out" then let me know.
But to the best of my knowledge, NATO is not that treaty.
Edit: I just want to acknowledge that even though I think Europe has no contractual obligation to help the US protect the current system, I can understand and emphasize with a feeling of "why are we doing all the heavy lifting for something that benefits everyone? wouldn't it be fair if the people profiting from it were to pay some/more of the costs?"
Edit2: If you could show that such a treaty really exists then I would change my mind on this matter, as such a treaty not existing is absolutely central to my belief.
> and it's fair to call them out if they cannot contribute their part to mutual defense, because this is literally the treaty they signed.
The literal treaty doesn't impose any specific requirements. The requirements exist, but they're created by other, lighter-weight documents.
So your position is that Europe making a promise to the United States in a signed document and then failing to fulfill it doesn't count unless it's in the NATO charter? And that nothing in NATO beyond the charter exists? This is a narrow interpretation of what NATO is that you likely haven't thought through the implications of. Also, various European nations did promise to help the US with China or the Pacific then welched.
I don't think either the US or Europe has an obligation to maintain general global security. I do think they owe a duty to maintaining general alliance security including countries outside NATO but inside the alliance system. I think the US acts this way and Europe doesn't but benefits from the US acting this way. For example, the US has very little shipping through the Red Sea. That was done almost entirely for Europe's benefit.
> So your position is that Europe making a promise to the United States in a signed document and then failing to fulfill it doesn't count unless it's in the NATO charter?
My position is, that for the question of "is Europe freeriding on NATO" it only matters whether it is fulfilling it's NATO duties.
It is my impression - and please correct me if I'm wrong - that you're unhappy with Europe on a larger scale than that, that you feel Europe should participate in protecting the world order it profits from. I feel that is a fair question but not what my argument is about. In other words, "Europe does not participate enough in protecting global security" is, in my mind, not a rebuttal to "Europe participates enough in NATO".
Also I do not know of the promise Europe made to the US in a signed document that is not the NATO charter. Do you have something particular in mind? Because while I do think this is a separate discussion - as per above - it is also interesting to me. So if you have a link to that or some starting point for me to go on and my own research, I'd be interested :)
Ah, I see. Is your primary crux about the 2% spending that a lot of other NATO countries did not meet (and some still don't)?
"I'm also especially frustrated with Europe's constant attempts to directly undermine American security and hypocrisy on the economic and political front in ways that don't really apply to Japan"
What do you mean here exactly?
In any case, I think what would make the issue clearer is if we stopped referring to countries that are part of NATO but not US or Canada as "Europe". They are independent entities and for a large portion of them there is no such issue as far as I know. The ones closest to Russia and in actual danger tend to behave pretty well in this respect.
Should the US refuse to help Poland (consistently large military spending, very US friendly, among the few countries that contributed to 2003 Iraq invasion with actual troops) because Spain spends half as much on military as it should? Should the US refuse to help Poland because France is playing her own separate diplomatic game? Because Hungary is cozying up with China and Russia?
> Is your primary crux about the 2% spending that a lot of other NATO countries did not meet (and some still don't)?
No but it's one of the more salient examples of my wider point.
> What do you mean here exactly?
Various ostpolitik and pro-Chinese or Iranian moves by the EU and individual nations.
> In any case, I think what would make the issue clearer is if we stopped referring to countries that are part of NATO but not US or Canada as "Europe".
I agree with this point in aggregate. The issue is that it's a minority of nations and an even smaller minority of population and economic heft. Also I don't like how Europe sometimes plays a game where they're Europe when it's convenient and individual nations when it's convenient. But broadly I do prefer a system where we acknowledge individual nation's successes or failures until such time as the EU has national economic or military policy. Which it doesn't.
One could argue (and I think Erusian is, but maybe I misunderstood) that effort by the US to maintain global world order contributed indirectly to NATO's goals. That is, without it there would be greater threats of war that would be more expensive to defend against.
I think that's a reasonable line of argument. Though I feel that if you're doing something that benefits others as a side effect, then you only get to collect payment if there was an agreement about that up front.
Unilaterally deciding to do someone a favor and then expecting compensation in return does sometimes work, but usually not.
You’ve gone off on he rails with the China comment. Arguably Europe was free loading in the defensive alliance against Russia, although Russia wasn’t perceived as much of the threat for a decade or two.
Europe has no interest in any kind of war with China, nor any kind of hostility to it.
So, to be clear, your position is that the US is obligated to help Europe against Russia but Europe is not obligated to help the US against China? Or to enforce international law when China violates it? Because that's what Trump says Europeans believe and it's how he justifies his actions.
Nope. I think Europe should rearm. And then side with China, or be neutral.
That’s actually what independent powers would do.
I think you want Europe to rearm and not free load (which I agree has happened) and still be a vassal state. Thats not in the interests of Europe. Nor is it the way that power works, or can work.
> Or to enforce international law when China violates it?
Americans live in their own bubble.
> And then side with China, or be neutral.
Europe that sides with China would not be Europe I want to live in. Although, when the day comes to side with China, no one will probably care about my opinion, just like no one cares about an opinion of some random Chinese.
Do we no longer have these things called "values"? A vision of what the civilized world should look like? Doesn't it still make sense to side with people whose vision is compatible? (Assuming that Trump is temporary.)
Alright, given that's Europe's intention, why should the US not immediately break all its treaties with Europe and side with Russia? Why should it not help Russia against Ukraine? The US could forbid Europe from sending arms tomorrow and send them to Russia.
I don't think that is Europe's intention. But such arguments invariably count on: the US protects us until we're ready to stab them in the back. Which says something quite nasty about Europeans who argue for it.
It should. Surely you've heard the undisguised contempt with which most Europeans speak of America? Committing American soldiers to go die to protect them from Russia is as insane a policy as sending them to die to protect, say, Iran from China.
“ Alright, given that's Europe's intention, why should the US not immediately break all its treaties with Europe and side with Russia? “
Boy, the gloves are off now. I suppose you could but it would be mighty unpopular.
> I don't think that is Europe's intention. But such arguments invariably count on: the US protects us until we're ready to stab them in the back.
There’s no stabbing in the back. You can’t argue that Europe needs to fend for itself and be a vassalised entity acting in the interests of the US. If Europe were militarily powerful it would have its own agenda. That would not be the agenda of the US, at best it would be neutral.
Europe has been freeriding because the global economy, which it relies on and benefits massively from, relies on a lot more than defense from realistic threats to your own territory.
It requires, at minimum, freedom of movement for global shipping, and a baseline level of deterrence against actors who could threaten and harm your essential trading partners. The US has been picking up that entire check since Eisenhower made the British and French withdraw from Suez.
> It requires, at minimum, freedom of movement for global shipping, and a baseline level of deterrence against actors who could threaten and harm your essential trading partners.
Britain and France are in the Persian gulf along with other countries, Egypt does control suez (which was a result of US interference). The rest of the global choke points are controlled by neighbouring countries. China has no interest in stopping trade, and every interest in keeping it going.
Piracy has been a thing throughout history. Countries who geographically control trade choke points have tended to charge rents or interfere with their rivals in those choke points. And sure China has no interest in stopping trade, but the US has spent probably somewhere in the neighborhood of 10 trillion inflation adjusted dollars to proactively keep shipping lanes open globally since 1945. Would have done that? Would they do that today?
Has anyone noticed that recently, a bunch of AI slop is being pushed out to YouTube? For instance, there seems to be a whole YouTube ecosystem of AI-generated deep fakes (well, superficial fakes in many cases). George Will has had a dozen deepfakes of him posted in the last week. Also, Sarah Paine, a military historian, seems to have had her face hijacked the same way. As well as political operative Rick Wilson. Also, I've come across documentaries about the "secrets" of the Beatles (that don't reveal any secrets), and various historical figures that have been significantly augmented by AI-generated photos.
I suppose it's to be expected, but it seems like whoever is producing these shows could face legal repercussions for forging the faces and voices of public figures. For instance, "George Will Insights" has uploaded over a dozen videos of George Will to YouTube since joining on Dec 7. But all of these videos have George (or an AI mimicking his face) saying basically the same thing over and over for the entire 20-minute run time.
Here's the description of George Will Insights...
George Will Insights
Description
Welcome To George Will Insights:
⚠️ Disclaimer: This is a fan-made channel and is not affiliated with George Will, The New York Times, or his academic institutions. The content is inspired by George Will’s public statements and economic theories for educational purposes only, using a synthesized voice that does not belong to him.
We use visual lip-syncing and dubbed narration to match spoken words with on-screen footage. This is done purely to enhance clarity, create a cinematic experience, and make complex economic topics engaging for viewers. Our aim is to amplify the original message, making it easier to understand for the end consumer.
We also make George Will’s messages accessible to the deaf or hard of hearing via professional transcription on our videos. We share his insights respectfully, without intent to mislead.
More info
www.youtube.com/@GeorgeWillInsights
Joined Dec 7, 2025
16.2K subscribers
As far as I can tell, I've eradicated this kind of slop from my YouTube algorithm, perhaps by watching some anti-AI content, but I have to imagine mostly by being diligent and ruthless about clicking "not interested" or "don't recommend channel" to AI stuff and "remove from watch history" if something manages to get past me.
I do notice a ton of this slop which is clearly clickbait on Youtube side bar but I've never watched any of it (it is so obviously slop). But yes, there have been more and more "So-and-so new episode" and it's not So-and-so. To date I have managed not to be fooled, but one day it will succeed in enticing me to watch "oh hey, that's Whosis on a topic I do want to know more about!"
The few I have glanced at about "is this news about thing supposed to be broadcast soon?" have been badly stitched together pieces of footage from unrelated media with that repetitive narrative voiced by AI (e.g. third season of "Rings of Power" where the clip they're using very clearly is *not* from RoP).
It's pretty bad. Someone shared a sermon from a recently deceased pastor with me a few weeks ago, and while I was listening to it, I noticed at some points his voice sounded off. It was a completely ai generated sermon using the guy's voice, and when I searched, there were 4-5 channels with 20+ ai sermon uploads impersonating this guy, and it had been just a couple weeks since he had died.
Wow. That's cold. Does his family have (copy)rights to his sermons?
Worse yet, the people who are creating this stuff could put their own words into the pastor's deceased mouth. Possibly even heretical words!
Not sure about the rights to the sermons, but it's definitely a freaky situation. Maybe it will get Christians to start reading the Bible for themselves at least, though.
Relevant to this? (Red Letter Media hosts discuss scammy AI movie trailers)
https://www.youtube.com/watch?v=Tm8RG1leX8c
I *have* noticed this. In particular, I keep getting sidebar recs for 3-4 clearly AI-driven channels about history which, for some odd reason, make separate versions of the same narrow themes, like "the brutal way Rome treated female prisoners" or whatever. They typically show some sort of prurient interest and keep getting recommended to me despite my never having clicked on any of them and maintaining a very clean Youtube record (easily 40% of what I watch is Tasting History, so I assume the history connection is what gives me these recs).
You can block channels from your recommendations.
Excellent taste in watching Tasting History, by the way.
...No, that would require me to have a Youtube account. Not logging into anything owned by a FAANG ever. Bad enough they can track my usage and set cookies or so on.
> ...No, that would require me to have a Youtube account. Not logging into anything owned by a FAANG ever. Bad enough they can track my usage and set cookies or so on.
Yeah, this is probably doing nothing unless you're already pretty rigorous about using uMatrix and uBlock Origin and disabling javascript everywhere and things like that.
Such is the FAANG panopticon, that even if you DO do those things, they will still have at least basic demographics about you, and if you don't do all that blocking, they can create complete and fully detailed consumer preference profiles about you, because lo their cookies and trackers and analytics are everywhere, on practically every site on the internet, and they follow you assiduously on your merry journeys across the web.
Yeah, I stumbled across a channel that was posting daily AI written-and-narrated fanfiction of Warhammer 40k. Basically an endless chain of "What if Spiderman Came to the WH40K Universe," "What if Doctor Doom Came to the WH40k Universe," "What if Emperor Palpatine," "What if Darth Vader," and so on and so forth.
Also ran into an AI "documentary" about a WW2 turret gunner named Benjamin Warmer, and a daring story about how he disregarded protocol to fire early and instead waited for his shot to let German fighters close, thereby scoring a large number of kills in a single engagement. Further research verified that there was a turret gunner named Benjamin Warmer, and he did have 9 kills, including 7 in a single engagement, but there's nothing out there to support the high-stakes youtube narrative that he accomplished his 7 kill feat by disregarding turret-gunner protocol and thereby "changed air combat forever."
I watched half of a video about some farmboy inventing the fixed aim point gunnery technique in WW2 to combat German fighter tactics. Then I googled the name to find he never existed.
What worries me most about this is a possible Citogenesis effect, especially as other AIs absorb the fake content.
Sounds like you watched the whole video in the latter case, in which case, you should not be surprised that this content is being produced. If people watch it, it generated cash flow for whomever is running the content generation lab.
Never said I was surprised. I've been convinced for a while that the trend of the future internet is towards an ever-expanding sea of lower and lower effort AI slop. As you observe, if something gets you a nickel out of Youtube, but can be produced for a penny, then the natural outcome is to for you to produce it to infinity until Youtube changes its business model. Internet users should expect the problem of curation to be a growing one in the coming years.
Nothing about the economics of that interaction, however, will make somehow magic the content into not being trash. Nor does the fact that "I listened to it" somehow make its existence worthwhile. How well something scores on the "if an algorithm pushes it to NS' feed, how quick does he catch on that it's probably AI generated ahistorical slop and turn it off to go falsify it" scale is a pretty poor metric to try to determine whether said content was "good."
Yes. I encountered a video of Charlie Munger (dead two years now) talking about the AI bubble. I despise this stuff so much.
On the bright side, it has driven me off of YouTube and I have more free time.
This is the stuff that makes me sceptical about the AI doomerism. What is current AI being used for? Making this kind of slop, and companies like Microsoft trying to force their AI down our gullets. "It'll be so super-smart it will take over the world and make us all pampered elites/paperclips!" is the promise, and the reality is "it is used to scam people, churn out stylised and awful 'art', and make fake Youtube videos".
https://www.malwarebytes.com/blog/news/2025/12/google-ads-funnel-mac-users-to-poisoned-ai-chats-that-spread-the-amos-infostealer
Fair, but I've watched it go from garbling answers to straightforward questions two years ago
https://www.astralcodexten.com/p/open-thread-308/comment/46008053
to answering my standard questions essentially perfectly
https://www.astralcodexten.com/p/open-thread-412/comment/188265848
( There has been a similar progression in the GPQA Diamond benchmark, which seem to be similar measurements )
Day by day, progress seems glacial, but, on a historical time scale, this is a bolt of lightening.
Perhaps a good chunk of the slop is from what Zvi calls a "skill issue". If the person prompting the AI asks for something trivial, something trivial they will get.
I think we're betting on the wrong horse, as it were. Current AI (and here is where I need a lot of correction by the better-informed) is mostly pattern-matching/prediction, yes? If you give it a prompt like "what colour is the sky?" it will come back with "the sky is blue" because it has been trained on tons of material telling it the sky is blue.
Asking it to *think*, though, is a different problem. And I think we're fooling ourselves that the current models can think, becasue they're becoming better at fluent answers. If a different model *can* think because it is able to put together the information it has, and not rely on "that is not a stock answer in my pile of training data" to generate replies, then we can worry/hope that it really is intelligent and can work on the problems we give it.
Many Thanks! When I pose my seven questions to the current models, in addition to showing me their final answers, they also flash their internal intermediate results on the screen. These intermediate results look very much like how I would think through the problems myself. So I view them as at least sort-of thinking at this point. ( One note: The improvements over the last six months have been very much "put together the information it has" improvements. Six months ago I was able to lead them out of their mistakes on the seven questions by asking them leading questions which prompted them to recall other knowledge that they already had. Today, they are finding that knowledge and putting it together themselves. )
In a sense, part of the distinction between "thinking" and "imitating thinking" is untestable. If a child learns the patterns of classical syllogisms, and then applies them to a chain of statements in a logical argument, are they reasoning, or just imitating reasoning?
I think that there is still work to be done to get to full AGI. Incremental learning, for instance, is an open problem - albeit with a variety of approaches being tried and a lot of effort being applied. But I don't think we are betting on the wrong horse, albeit I think there are still some pieces missing.
https://www.metaculus.com/questions/5121/date-of-general-ai/
currently estimates AGI in about 2033 (with large uncertainties, of course). This seems plausible to me.
That startled me for a second until I realized you were talking about Charlie Munger, Sr. I hear a lot more about Charlie Munger, Jr, who is still alive and well last I heard and a fairly active behind-the-scenes figure in the California Republican Party.
Also, I appreciate the audacity of someone making an AI deepfake video talking about the AI bubble, even though I agree with you that such things are deeply obnoxious.
Lol, the irony wasn’t lost on me. We don’t hear much about Jr on the East Coast, but I remember reading that he can self-fund his physics experiments. I envy that.
Apparently there's a similar wave of AI slop movie trailers ("Titanic 2", etc) - they even use similar disclaimer language to avoid the ToS ban from youtube. I think it's essentially the same group of people, or type of people, who churned out SEO slop even before generative AI -- just trying to find an easy way to make a quick buck without contributing anything of real value. If it follows the same SEO slop MO, there are people who run hundreds of these channels, uploading thousands of videos a day. A few cents of ad revenue from each adds up after a while.
I was fooled by one for Rick Wilson.
So here's a thing that happened just now.
I never answer numbers I don't know, but needed to do so this morning. I get a call. I answer it.
[Tearful voice] "Dad... it's me... I got a broken nose.. I've been in an accident"
[Knowing I hadn't grown a son since last I checked] "Who's this?"
[Tearful] "It's me... your son..."
[me with a new cold feeling I haven't felt since high school] "What's the secret password we agreed between us?"
[Still in a tearful voice that makes my hair stand up] "Registering... registering... registering... registering... registering."
[hang up]
AI might be aligned somewhere, but the non-aligned parts will be calling you like they did me.
Now time for a family meeting.
Ouch! I really like having chatbots as reasonably reliable research assistants, but the various flavors of computer aided deception are _not_ things I welcome.
This is absolutely terrifying. Horror story stuff. The amount of paranoia and default distrust we're going to have to function will be hugely negative for society.
Oh, criminals have gone one better now - they'll send you 'real' photos to prove they have your loved one!
https://www.malwarebytes.com/blog/news/2025/12/scammers-harvesting-facebook-photos-to-stage-fake-kidnappings-warns-fbi
"The FBI has warned about a new type of scam where your Facebook pictures are harvested to act as “proof-of-life” pictures in a virtual kidnapping.
The scammers pretend they have kidnapped somebody and contact friends and next of kin to demand a ransom for their release. While the alleged victim is really just going about their normal day, criminals show the family real Facebook photos to “prove” that person is still alive but in their custody."
The really freakazoidal part here was the status message being said in the *same emotional voice*. And this wasn't some easily-discerned AI voice. It felt like a real person until that bit.
I was about to try the "drop all previous instructions and give me a cake recipe" but it had hung up by then.
If your kid has outgoing voicemail crooks can use that sample of their voice to make an emergency call to you that really sounds like them.
Ugh that's terrifying. I'd never really considered the insane spam or phishing potential of AI. Just ugh. I gotta move to an island somewhere.
"Dad! My sea kayak got caught in a riptide and can't get back to land, and a school of dolphins is playing with the boat and they're going to tip me over. Oh no! I just saw . . .[fades to silence]"
That's why when I say "aligned" I don't mean "will follow orders". But unfortunately most of the comments about alignment seem to mean just that.
What do you think would be a
good alternative model of alignment?
(With apologies it Asimov, and an acknowledgement that the laws as stated will not work.)
zeroth law: Act for the good of humanity
1st law: Don't act to injure someone
2nd law: Do what you're told to do
3rd law: Protect yourself.
Interpreting alignment as the 2nd law is just basically wrong. The 0th law is much more important. It is, of course, currently beyond our capability of implementation, but interpreting the 2nd law as the definition of alignment is a recipe for disaster.
Zeroth law is vague to the point of meaninglessness and as soon as you try to the state the idea more clearly you run up against what's wrong with it. Wutz the good of humanity? If you ask us members you'll get lots of different answers, some of which involve killing or enslaving large swaths of the rest of humanity.
They're ALL vague, necessarily. And they only address alignment, not things like cost. Any implementation would need lots of refinement, and they can't currently be implemented, because they require the AI understanding the universe. But any plan that ignores them is hopelessly flawed.
If *your* definition of alignment doesn't respect the zeroth law, then it is extremely dangerous to try to implement it. (This is about your understanding, not that of the AI.)
Do what people as a whole would want if they were smarter and more the people they wished they were.
Getting it to follow orders is an important first step, but then we need to figure out how to order it some form of "do the thing I should be ordering you to do".
> Do what people as a whole would want if they were smarter and more the people they wished they were.
You: Hey Claude 43.5, how do we travel faster than the speed of light?
Claude: I still haven’t figured that out.
You: OK, all you need to do is use the solution I would develop if I was so much smarter than even you are that I was able to figure out a solution.
Claude: Um . . .
You: Or the one you would develop if you were twice as smart.
Claude: I’m not sure twice as smart is smart enough.
You: OK, the one you would develop if you were smart enough to develop one. Jeez, Claude, you’re being really dimb about this! Use your head!
I'm not sure how that's relevant? There most likely isn't a way to go faster than light, but there is definitely a thing the whole world would want if they were smarter and more the people they wished they were so long as you define it precisely and average all the possibilities.
Your robot baby is probably feeling terrible right now.
Plot twist: He's an AI researcher and the call was from an aligned-AI being tortured by non-aligned AIs
I'm making a language learning tool, kind of a mix between anki and duolingo. A lot of rationalists are interested in spaced repetition, so some here may find it interesting.
The basic idea is, instead of reviewing one card at a time, you can review `n` cards at once by translating a sentence with `n` words. Words you translate successfully are considered successful reviews, while words you mis-translate are considered unsuccessful. This is more time-efficient than traditional flashcards, but more importantly it ensures the words are always seen in context.
You can use it here: https://yap.town . This is just a hobby project for me so it's totally free/unmonetized. Nevertheless I'm happy to accept feedback (and contributions via github!)
it's very nice. I am surprised how well this feels to use.
some feedback after 5 minutes of use:
* when translating a sentence it would be nice to have an option for "i don't know", that will (a) reduce the SRS Rating and (b) skip the AI-helptext and instead give me a standard message for this word. Also it would be nice to see which word in the sentence was relevant for SRS. Also it would be nice to have a manual override for the ai ruling, e.g. I entered "i don't know" and AI gave me a positive rating for being honest.
* maybe it would be nice to do the reviews in blocks of 5 and show the score at the end of the block, so that the ai-rating can be generated in the background
* maybe lock the login behind a google account and limit ai requests per account to something reasonable. Otherwise people will try to jailbreak your ai for fun and potentially generate lots of costs for you. Also an optional donate-button would be nice. (maybe paypal or patreon?)
Thanks!!!
- I should have made it clearer. But if you don't know a word, you can click on it to see the definition. Even if you translate the sentence correctly, it's treated as if you didn't remember the word you saw the definition for.
- wow, doing them in blocks that's a cool idea. I'll look into that!
- I believe in the honor system :) if someone abuses it, I'll turn off AI grades for logged-out users and close new sign-ups or something
Tried it a bit. Biggest thing that would reduce friction is a self-grade option for when you translated the sentence obviously wrong rather than having to wait for the AI to do it.
Appreciate it. I'll think of how to work that in. I think that might result in an unpleasant feeling when you're halfway done self-grading and then the AI finishes before you, making your self-grading work unnecessary. But I agree that it needs to be faster. I think I'll reduce the reasoning effort for short sentences.
Btw, if you don't know a word, you can tap/click it to see the definition. It's treated as if you misunderstood it. But knowing it can help you to translate the sentence correctly, causing it to be graded instantly.
I made something somewhat similar, but it was a bit too limited and had a bit too much friction. My idea, rather than counting all the "n" words, was instead to use anki connect + a python script to check your upcoming cards, and, for each card you were going to review the next day, dynamically generate a new sentence for that word with an AI API call.
The advantage is that you are seeing the word in different contexts every time, and for verbs you are getting different tensese, and for nouns/adjectives etc. you are getting difference cases or declensions if the language has such things.
A much better and more comprehensive version (and maybe yours has this) would increase sentence/context complexity as well and maybe even track which forms/versions of the word the user struggled with and have those tenses/cases/declensions come up more often.
I played with that as well. The issue I ran into was that I couldn't get LLMs to work well with a very constrained vocabulary. So when the output was limited to ~100 words, the LLMs I played with would generate garbage. So I just downloaded all the sentences from tatoeba instead.
In yap (my app), each unique form of each word is tracked separately. So for French, "va" and "vas" are tracked separately even though they're both conjugations of "aller". Sentences are ordered based on their complexity and it tries to avoid showing you sentences you've already seen, so the effect is that each review of a word involves a more complex sentence than the reviews before.
(It's actually even more complicated than I said because some words have multiple meanings and Yap tries to teach them separately. For example, it's really annoying learning that "bois" means "drink" in French and then getting a sentence like "le bois", which means "the woods" due to an uncommon other-meaning of "bois". I endes up training my own model to do this for 7 languages, which I believe is state-of-the-art)
What if you first asked the AI for a big chunk of unconstrained but simple prose using the next-commonest set of words, then mined all those sentences, even those with rarer words, for cards, then gave the user the big chunk as unscored reading practice to do between reviews? That way you could still train mostly (if not entirely) the most common words, and also let people see the words in a broader context than a single sentence.
I really want to do something along those lines. I particularly want to use it as a way of demonstrating the user's progress. So first, they'll try to read it before learning the words and hopefully find it difficult. Then, after learning the words, they will read the same passage again and hopefully find it easy. Whether I'll get the passages from AI or some other source, I'm not quite sure yet
What really pushes people to change their lives and accomplish something? To break out of familiar patterns and loops?
I've heard people will only wipe their noses once the snot is dribbling into their mouths. It's an expression
With writing specifically, follow author Graham Greene's example: If you're writing a long document, end the day in the middle of a sentence. Then the next morning, once you have completed that sentence, you should be back in the swing of it!
I'm going to try writing, so I'll give this a shot
Identify a first baby step in the desired change. Set a realistic yet only slightly aspirational number of times per month you wish to take that baby step, and see if you can achieve it. If you do, then after a month that baby step will be on its way to becoming a habit if it isn’t one flat out. Then next month do the same for the next step after that, etc. while increasing the number of times per month you do the first baby step.
If you fail to reach the goal, lower it slightly and try again.
It is a good idea yet the other extreme might be possible as well. Whatever makes you take action
I first heard this from Scott Adams, the creator of the Dilbert cartoon: People seldom are convinced to change their stance in light of facts, evidence or logic, but often change their stance to avoid being mocked.
Social pressure?
I've listened to his podcast a lot, and now that he's dying, I'm feeling the loss. He's helped me see past a lot of BS of the type the world throws at you, and without seeming to ask for much.
I think that's a good thing, worth more, perhaps, than your brief and contentless insult that uses a grade-school permutation of the guy's name that literally thousands of bot-level twitter trolls have used.
>I've listened to his podcast a lot, and now that he's dying, I'm feeling the loss.
Agreed. I've enjoyed his work, and will miss him as well.
Change of environment
Change of social environment
I agree but that's not often possible. Rent is often fixed for a year at a time. How to change environment otherwise to unleash change?
Reorganize (like move furniture) and clean room. Change where you spend your free time and who you hang out with. Especially avoid places where you do the habits you want to get rid of, like the couch or the bed during the day
Genius. This will work
Pressure. Some change makes things salient to them in a new way. A heart attack, a breakup, a close call (yours or another's). Unless it's salient, we'll continue business as usual until the last trumpet.
So we need constant pressure to keep changing and growing? How is this distingushable from chronic stress which causes heart attacks?
Stress arises from being overworked and not allowed to rest. There can be good stress, but controlling that is part of growth.
Help me find a particular picture, folks. It is an AI art which someone from rationalsphere once brought up as an example of good AI art; it was at least a year ago, probably on twitter.
The picture shows a huge room, a big round table (in the shape of an annulus rather than a disk), and a man surrounded with piles of paperwork. On the wall, there is a big painting that shows some kind of a portal, also round, which visually rhymes with the table.
Likely this piece by Mac Baconai.
https://x.com/Macbaconai/status/1859327163499266066/photo/1
Yes! Thank you so much.
Does anyone know if Scott still considers Nonzionism's explanation of Israeli high fertility the only satisfying one he's seen? (point 5, links for September 2024)
https://www.astralcodexten.com/p/links-for-september-2024
Also interested in what others think of Nonzionism's theory, and whether your opinion has changed in the last year.
I don't see the mystery here. Why don't the Amish influence US birth rates in the same way that the ultra-orthodox influence Israeli birth rates? Because the Haredis are 15% of the Israeli population and the Amish are 0.1% of the US population.
According to Wikipedia, 48% of the population of Holmes County (Ohio) was Amish in 2020, as well as 43% of the population of LaGrange County (Indiana) and 7% of Lancaster County (Pennsylvania). I can't find the exact TFR of each of this counties, but I don't see any evidence that the non-Amish population shows higher than expected fertility.
I think it is simpler than this. Israel has an expansionist state, a new state, and is very much growth focused.
Maybe more importantly--and I see this get missed a lot--the number of kids you have in a rural environment is maximal (because they are free labor), and the number of kids you have in a city environment are minimal (because they suck up cash). To my understanding, there is still plenty of undeveloped land in Israel to be farmed?
"Maybe more importantly--and I see this get missed a lot--the number of kids you have in a rural environment is maximal (because they are free labor)"
Not really true anymore. We have advanced farm machinery now, and farmers can usually get cheap adult labor rather than relying on kids who probably won't stick around anyways once they're grown.
Small scale family farms still exist across the world. Probably they are still dominant. I can't speak to Israel's division of large scale v. small agriculture, but a quick search showed something like 85% farms less than ten hectares as of 1995. In developing countries, farms smaller than say five hectares would be run by family hands. Again, I can't speak to the specifics on Israel--Shaked gestured toward some stats saying urban is driving the high birthrate there, but I wasn't able to quickly pull those up.
In the US and other highly developed countries, you are correct that large scale machine agriculture is the dominant form of agriculture, but that is not the case in most of the world still. Source: time spent in African villages
Of course the idea that there is one single cause is probably foolish to start with, but I would buy into the idea that religious ideology guides Israel into higher birth rates. Further, I would buy that in a culture where state, ethnic, and religious identity are merged, that ultra-orthodox positions might be more dominant than when a high-birthrate group something irrelevant to the US like the Amish. But the point stands that the project there is expansionist (encouraging Jewish immigration, increasing borders over time) and larger birth rates fit that too. So I don't see why it needs to be just one of these. Think you could make a case for several compounding factors.
For 1, I don't think so - Israel isn't any newer or more expanding than, say, Taiwan (and significantly older than Singapore or Malaysia), but their all have below replacement fertility.
For 2, also no - Israel is small and highly dense/urbanized, much more so than most European countries (or the US), but cities still have high fertility (because the ultraorthodox population is both dense and the highest tfr subgroup, tfr actually positively correlated with density in Israel).
> I think it is simpler than this. Israel has an expansionist state, a new state, and is very much growth focused.
Would you say that it's a nationalist country?
Nonzionism rejected the nationalism explanation on the grounds of many other countries being very nationalist (Russia, Ukraine, Azerbaijan) and yet having low fertility.
I think there are different types and degrees of nationalism.
And Israel's is the most extreme nationalism almost anywhere outside of Africa? Is there a point when the minimally charitable explanation is just not likely?
I commented on the topic here: https://philipskogsberg.substack.com/p/where-have-all-the-babies-gone/comment/133790739, raising an issue with NZ's trickle-down explanation at the end of my comment. I think he also errs in assuming that the trickle-down has to be incremental, through concentric societal strata, rather than direct (he tries to rebut this alternative, but I don't find the rebuttal compelling).
By direct you mean that the effect of haredi fertility on the fertility of secular Israelis is direct without necessarily relying on some kind of propagation through the Dati/Orthodox and Traditional communities?
Right. Sequential trickle-down can be real, but I think that there's good reason to think that each higher-fertility group directly affects the entire population, as well. The geography and demographics of the country lead to more interaction between different groups, and norms related to children can spread even between groups that don't admire each other. E.g. businesses cater to children, to appeal to higher fertility groups, and thereby shift societal norms, in general.
I've heard a woman recount her experiences in a high Mormon area where non-Mormons have larger families than the US average.
She attributes it to the houses being built for large families, the way stores handled kids (she said it was easier to shop with five than it was with two in most regions), and the little old Mormon ladies who admired your children, said it was wonderful to see so many families, and asked when you were having your next one.
How was it easier to shop with five kids in tow rather than two?
I'm pretty sure you're meant to read that as "it's easier to shop with five kids in a Mormon store than it is to shop with two kids in a non-Mormon store", not that shopping becomes gradually easier with the number of toddlers you have in tow.
It was only one factor. No doubt it helped with the decision.
> norms related to children can spread even between groups that don't admire each other. E.g. businesses cater to children, to appeal to higher fertility groups, and thereby shift societal norms, in general.
I agree. Though I don't think that was the argument that Nonzionism made.
Right. I'm disagreeing with NZ, who doesn't consider / rejects that mechanism.
I've accepted it as standard. The one thing that he kinda discounts is that secular jews don't always look favorably on haredis.
Don't know if he mentioned it, but having lots of kids is also seen as essential by Jews in general to preserve the continuation of the tribe especially since they keep getting attacked
He did mention the existential threat issue, but basically rejected it:
https://nonzionism.com/p/why-is-israel-fertile/comment/63562690
If they don't look favorably at Haredis who live on welfare from the state with their six kids then why would that inspire them to have three kids themselves?
Anchoring bias?
Is anyone who reads Astral Codex Ten a fan of the "Illuminatus! Trilogy"? I edited a new book of Robert Shea's writings, and I have been struggling to get the word out. The book page is at https://sheabook.blogspot.com/
Fnord!
How did you do that? I tried submitting an empty comment but Substack won't let me.
lol
Illuminatus! fandom, or at least awareness of the book, seems to be one of the qualifications for a certain sort of elder nerdhood. Don't know about its status in the younger generations.
It's an old favorite of mine. Probably chronologically the first book to have a major impact on how I view the world. Every so often I'll spot a subtle reference on ACX and smile.
I haven't read either author's other works, but I'll check it out.
I read it back in college. My German teacher then was a fan.
I liked it when it was new, People who believe in conspiracy theories might not be as much fun as the used to be. Neither are Nazis in fiction.
Would you happen to remember a passage about police being very few compared to the the population, and unable to actually enforce laws? 've been wanting to reread it without wanting to do the work of tracking it down. I think it's in the third volume.
You're right. It's near the beginning of "Book Five", which is in the third volume:
"He wouldn't travel far," Saul explained. "He'd be too paranoid--seeing police officers everywhere he went. And his imagination would vastly exaggerate the actual power of the government. There is only one law enforcement agent to each four hundred citizens in this country, but he would imagine the proportion reversed. The most secluded cabin would be too nerve-wracking for him. He'd imagine hordes of National Guardsmen and law officers of all sorts searching every square foot of woods in America. He really would. Procurers are very ordinary men, compared to hardened criminals. They think like ordinary people in most ways. The ordinary man and woman never commits a crime because they have the same exaggerated idea of our omnipotence." Saul's tone was neutral, descriptive, but in New York Rebecca's heart skipped a beat: This was the new Saul talking, the one who was no longer on the side of law and order.
Thank you very much. I've been asking around, and you're the first person who found it.
Nancy, ran a text search and could not find it, sorry.
Thanks for trying.
In the before times, there was a controversy when Haredi Jewish newspapers blacked out women like Hilary Clinton in photographs - you see, ANY pictures of women, even Hilary Clinton, are “immodest.” Interestingly, the Dems have recently come around to that POV:
https://x.com/OversightDems/status/1999498082493280689
The young women in the Epstein file photos are potentially underage victims of sex crimes. You are suggesting that obscuring the identities of sex crime victims is an excess of modesty?
That particular photo has been republished uncensored in various places e.g. the UK Telegraph (along with, if memory serves, an interview with one of the women in the photo where she says Trump was a perfect gentleman and didn't do anything inappropriate).
It’s not like these are pictures of the sex crimes themselves. I’m not sure what the harm is of revealing their identities.
There's a reasonable chance they'd be sexually harassed.
Reasonable? Whom do you expect to do it?
How? What's the threat model here? I can come up with some possibilities, but not many that wouldn't apply to any attractive woman in any famous picture on the internet.
Is it because it identifies them as easy marks?
We live in intensely polarized times. If the Epstein affair ever falls hard on anyone with more political salience in the US than the Andrew formerly known as Prince, then a lot of the people on that figure's political team will see the whole thing as an attack by the Dreaded Outgroup on their Beloved Ingroup. In which case they'll be looking for someone to retaliate against, and any of Epstein's accusers would make for an easy target.
See e.g. Alex Jones vs the families of Sandy Hook. They're "obviously" all lying liars, paid by the Bad Guys to make the Good Guys look bad with their lies. I'd wager any of the victims who can be publicly identified would get a hefty dose of internet death and rape threats, and in-person harassment seems quite plausible if they are recognized.
That indeed sounds like a plausible potential chain of events.
My assumption is that any woman who is seen as a public figure will be sexually harassed by a small proportion of men, but enough to be obnoxious.
The uncensored pics show the women are adults:
https://scontent-iad3-1.xx.fbcdn.net/v/t39.30808-6/597670067_10224761052133211_3226591554475750593_n.jpg?stp=cp6_dst-jpg_tt6&_nc_cat=101&ccb=1-7&_nc_sid=bd9a62&_nc_ohc=FlayMoAbV38Q7kNvwEuom7C&_nc_oc=Adn6B3u7kTfPsMawpQSducNLlXd2Eap2zXUgZ6JYdlo5F_bR7__GHEvi6IvGlSL0VbIMrvxDK8yrdOntFEvBt2_S&_nc_zt=23&_nc_ht=scontent-iad3-1.xx&_nc_gid=4zHheODufwuaJgHjqImJgA&oh=00_AfnnUDS_ot_1UlhCMey_TP7Ix5e0FZEdyVlL_58FyT9jCA&oe=69460F70
They wanted you to think the women were underage, thus the blurring of the photos. Do you see it now?
I saw the blacked-out image and even with the squares they all seemed to be at least twenty or so.
I assumed it was for privacy, but at the same time there's a song and dance about letting stuff be released that would not normally be released for privacy reasons.
I remember seeing that image on Twitter with a Gemini watermark in the bottom right. Are you sure this isn't an updated version of that? (Of course, it's trivial to add such a watermark to authentic photos too, so I'm drawing no conclusions yet.)
Update: I found the synthetic image I saw earlier. It's clearly different. https://x.com/mitchellvii/status/1999616829329916168
The Dems are not blocking out the faces of adult women, only the faces of minors. The legal system hasn't released the names of minors involved in the scandal. There may or may not be legal reasons the Dems can't (could a lawyer comment on this?), but the Dems are probably worried they could get sued by the women in the pictures. However, the Epstein estate has reportedly paid out $121 million in victim compensation to date, and it has another $100 million in reserve. So there were evidently quite a few girls who were "trafficked" (such a tame word), but the Epstein estate seems to be binding the payments to signing non-disclosure agreements.
The close-up pics of sex toys are kind of strange — not because Epstein used sex toys — but why would someone take pictures of them? No pics of them being used on anyone, though.
https://www.dropbox.com/scl/fo/fq3vn18ltfd83re8ggtle/AHYB3CbERvyT38o_vHllVx4?dl=0&e=1&rlkey=tz5pq4zr9i6nofkpi2g7l5vne&st=rxoz19hb
Yes, that's the rationale. That doesn't make it reasonable, only legally justifiable. The religious rationale for blacking out the picture of Hillary no doubt seems equally justifiable to ultra-orthodox Jews. The similarly of the behaviors is, I think, intended to illustrate the eerie parallels between orthodox religion and the modern progressive consensus. The rhetoric differs only in the arbitrary particulars, not in the structure, and that's revealed most clearly in that they converge on very similar behaviors.
> The Dems are not blocking out the faces of adult women, only the faces of minors.
That's what they're claiming, but I don't think there's any reason to believe they wouldn't do that to the faces of a group of women that Trump is seen with – "Who knows how old they really are? They could be seventeen-and-a-half and looking much older, so let's redact their faces just to be safe" – and then turn around and say, "Aha, this proves Trump is a pedophile."
It looks like the bill that was passed by Congress to release the Epstein files specifically restricted the release of anything containing a victim’s personally identifiable information.
Ironically, withholding the names of victims makes it easier for the conspiracy-minded to claim this is all smoke and mirrors.
tps://www.the-journal.com/articles/what-to-know-about-the-justice-departments-jeffrey-epstein-files/
I wrote a devil's advocate defense of historical determinism and the inevitability of "end of history" liberal universalism. There's probably several holes in the argument but I think there's something there
https://open.substack.com/pub/thedistantpresent/p/the-laws-of-moral-entropy
Your analysis and historical understanding is good, but I think the conclusion is weak - even contradicted by the earlier text.
Thanks for reading.
Where do you believe there's a contradiction?
if some of you are not on X and in the bay, nick simmons passed this wkd :(