quiet_NaN
3hr ago
Suppose a time traveler came back from the year 2100 and told you that, in his time, Britain (or Germany, or France) was 99% white and 94% Christian, the way it was in 1951. Unfortunately, before he could explain how such a thing came to pass, he gets hit by a lorry, dying instantly.
Suppose a time traveler comes to 1920 and tells people that in 1944, there are hardly any Jews left in Poland. What do you think had happened? Do you think the gentile Poles (or even the Germans) thrived?
Generally, societies which are willing to enact ethnic purges are also shitty in a lot of other ways. You do not get to 99% white simply by removing the recent immigrants, you would need to round up Asians whose ancestors have lived here since 1970.
Likewise, unless Jesus starts granting his followers spell slots, I do not think that 94% Christianity would happen by organically. My read on the 94% is that in 1951, parishes filled important social functions and most people who had doubts about theism found it wiser to stay in the closet. It is hard to make parishes great again, but easy enough to drive the non-theists back into the closet, Cromwell could do that.
Repressive regimes do not go away simply because they have arrived at an ethnically homogeneous society. Any Britain which will go back to 99% white and 94% Christian will be orders of magnitude more terrible than 1951 was.
quiet_NaN
14hr ago
But there is a whole treaty there preventing such things, the ECHR. Which is from 1950's and that makes me think, if those sort of treaties existed at that time, it wouldn't be unreasonable to say, since we are further left now than in the 50s, that we are now full blown communists, or at least, the policymakers are.
I do not get this argument. The ECHR prevents you from directly deporting someone who makes a claim of asylum. Or summarily executing them for entering your country illegally. When the UK signed that treaty, they were very much not "full blown communists". How does the continued adherence to that treaty make them commie today?
Practically, the ECHR does not say you can not deport anyone ever. Even us Germans are deporting people ... to fucking Taliban Afghanistan no less.
--
Of course, mid-2026, being worried about immigration feels like being worried about palace intrigues in Constantinople in 1453. I am a software developer, so I am first in line for getting replaced by the machines, but I certainly will not be the last. All white collar work is up for grabs. Blue collar work might still be around for a decade (present rate, no singularity and all that), but even before demand might crash. Unlike the PMC, the data centers will not need a lot of people to fix their cars.
Against this background of "the economic value of the median citizen will be marginal to nil in the near future", sprinkled with x-risk and visions of God Emperor Altman to taste, "OMG whites will become a minority" does not sound like an urgent issue, to put it lightly. Frankly, I observe that the people who get elected on anti-immigration platforms are mostly buffoons and grifters and rarely wise statesmen. History might not judge Trump by the spectacles he holds on the lawn of the White House or in Iran, and more on how he dealt with AI.
(I realize that this is a fully general counterargument. "How can you care about $X when literally $Y is happening right now.")
quiet_NaN
17hr ago
Here is a linkedin-free writeup. Low entropy key generation is an oldie but a goodie -- I recall when ssh keys generated with Debian had this problem.
For this particular story, I think that the morale is that rather than worrying about increasingly far-fetched modes of attack ("yes, you can keep your key on an airgapped encrypted OpenBSD laptop, but what if someone steals that laptop and manages to brute force the passphrase, would you not prefer some hardware encryption where the chip clears itself after three wrong passphrase entries?"), you should worry about "what if my niche crypto implementation is not as secure as anticipated?"
Also, if you make hardware wallets, auditing that the crypto is implemented correctly seems to be (not to put too fine a point to it) THE ONE JOB YOU HAD? Most failing cryptography fails not because the math is unsound, nor because someone has a quantum computer with sufficient memory, but simply because implementations are hard and people write buggy code. I know enough about crypto to not try to roll out my own implementations. "Key generated from low entropy PRNG" is a failure mode we had before. I recall Truecrypt simply prompting you to type a lot of noise simply because they did not wholly trust their supposed hardware randomness sources.
The #ifndef vs #if thing is likewise fun -- sure, usually it is not that harmful.
I will go out on a limb and say that this seems exactly the kind of bug an enterprising cybercriminal would discover with the help of an unfettered LLM these days. Being a hardware wallet where all the paranoid people park their coins makes it a juicy target, of course.
quiet_NaN
2d ago
At the end of the day, a majority believes in Newtonian Ethics. People may mostly be indifferent to people dying somewhere in Africa, but they tend to care if the same people try to reach Europe and bother us by drowning close to our shores.
quiet_NaN
3d ago
Agreed, it seems like it was conquered in 1415. If Ceuta is a colonial holding, then so are Wales or Istanbul.
I mean, it is certainly an overseas holding, they would not have kept their independence from Morocco without support from Iberia, but that does not make them a colony.
quiet_NaN
3d ago
The biggest news of the week is the ongoing Moroccan invasion of Ceuta, Spain. Ceuta is not on the European mainland but on the North African continent. Most recent estimates put about 60,000 migrants having crossed into Spain's North African territory, and the footage of the invasion is catching worldwide attention.
TIL Ceuta.
Of course, our asylum rules are clear: once they are in Spain they can try to claim asylum, and it will be up to the Spanish government to house and feed them while they are processing the claims, and then deport the ones who do not qualify back to their country of origin (which is presumably not Morocco). These are the same rules which allow us to tell women from Afghanistan that they can not claim asylum within the EU while being outside of it, and also penalizing airlines who transport people without visa to EU soil, with the clear hope that they will give up on trying to come here or at least drown in the process.
Given footage of Morocco's Royal Gendarmerie standing by while truckloads of men unload and rush the border, and the lack of any apparent effort by Morocco to prevent or control the situation, the situation seems to have been coordinated with Pedro Sanchez describing the situation as an "attack."
It is generally not the job of the police of other countries to keep refugees out of your borders, if you care about that kind of thing, you should install fences.
The situation would be different if they had allowed them to launch rockets or shells into Ceuta, because defending from these is pretty hard, so a government which allows such attacks to happen is failing in their international duties. By contrast, intercepting people crossing into your country has been doable since we invented agriculture.
I mean, I imagine the EU has been bribing most countries near its borders to house (or imprison) refugees rather than let them move on towards Europe, but that is at worst a breach of contract.
quiet_NaN
4d ago
First, it seems to me that there was little in the way of chattel slavery in England before colonization. There were serfs, sure, but they were generally if barely considered as human beings, being Christian and all. Chattel slavery is a different category from that.
The other thing is that most nations did not set out to become a nation which did things better than existing nations did, the US is the exception here. "Slavery was normal" works equally well as an excuse for Leopold II as it works for the US. Most apologists of the US argue (often convincingly) that the system (freedom, democracy, rights) is what set the US apart from contemporaries. "Hey, we were no worse than Leo II" is not a great way to plead exceptionalism.
quiet_NaN
4d ago
It is not just the nukes. Around WW1, large scale warfare between top tier nations simply became maladaptive. It used to be that being good at -- and eager for -- war made a nation thrive. These days, it is very much the opposite, at lot of countries would have a better quality of life if they overcame their willingness to spill blood.
Though to be fair, from what I read, the author does not bemoan the lack of large scale conflicts, and this is just commentary by the poster.
quiet_NaN
4d ago
The Remington Bolt Action numbers don't match up at a basics physics level: they give a 150 grain (9.7 gram) bullet at 812 m/s and 247.6 m/s as having 2777 J and 258 J energy, respectively. If the energy measurements are right, that has to be a 130 grain bullet. If the weight is correct, the energies were 3204 J and 297 J.
To be fair, the relationship between energy and velocity is quadratic. That is like eighth grade math, man. And obviously it should not matter because they have surely written down the exact provenance of any bullet they fired down to the lot number, right?
Also, a study with group sample sizes of N=1 and N=2 surely is even an embarrassment among the medical community? And that is before they arbitrarily (but non-randomly!) grouped the tests into the 150mm and 250mm thickness categories.
Also, this is physics. Only a complete idiot would make a bunch of measurements of the time objects need to fall to the ground and their height from which they were dropped, fit a linear regression curve through them, declare victory (significant!) and publish their coefficient of regression with a p value. Instead, what any high school kid in physics class would be expected to do is to find the value of gravitational acceleration which best explains their data. From a junior physics student, I would furthermore expect an exhaustive error analysis.
I am in particle physics. Obviously there are tools which can well describe the interactions of particles of all sorts with matter, be it a proton flying through your detector or a photon in the interstellar medium. The same is very likely true for other niche interests, like how buildings behave in earthquakes -- 99% of the population might not care (beforehand, at least), but the few people who do will likely make a decent job to model reality.
I do not see why small arms should be any different. To be honest, it does not seem all that difficult. Soft tissue is mostly water, so I would have assumed you could learn a lot about bullet penetration simply from buying the cheapest vegan hydrogel and firing bullets into it. Furthermore, it kinda seems a thing people would care about, given that militaries all over the world are still operating rifles. The idea that they would care about the cavitation effects of a nautical propeller but not about the interaction between humans and bullets seems absurd.
Even if water is too simple a model, I assume that most of the drag on a bullet does not come from the human soul. A semi-empirical model of wound ballistics based on animal studies seems reachable. You don't even have to use live animals for once!
That musket ball thing is just the icing on the cake. I am very much not a gun nut, but I would rather have a 5.56mm NATO go through my arm than a musket ball. From a humanitarian perspective, the redeeming qualities of the musket were that it was inaccurate, had a limited reach and took a long time to reload compared to latter battlefield horrors, but I imagine getting hit by one was gruesome.
As a side note, while I am very much not a domain expert, measuring the pressure exerted by a bullet seems particularly painful. Just measure the loss in velocity (time of flight measurements at those speeds seem easy, the offset of penetration between two rotating paper disks will do if you don't want to buy a Doppler radar or fancy light barriers). From the force which acted on the bullet in the gel you can use computer models to estimate what the pressure at the tip of the bullet actually was, but in the end you will want to use computer models to investigate hydrostatic shock in any case.
From your description, this smells like cargo-cult science, but I feel that this is not a field which has any excuse to be cargo-cult science.
quiet_NaN
5d ago
LLM text written for humans can be very useful. For quick research, short of having a team of experts on retainer who are obliged to answer you within a minute or two if you text them, the bots are obviously the answer. If I am debugging a specific Linux problem, I am not going to care that Scott could have presented the answer in a better writing style.
Here is a CW analogy for you. Guns are undeniably useful. Cheap. Scale well. Incapacitate physically stronger opponents without ever getting into close combat. Kill your enemies without needing to deal with the vagueness of poisons or henchmen.
Yet inexplicably, many places and societies restrict the use and sometimes even the bearing of guns. There are whole sports built around the fiction that you can't just shoot your opponent. How does this fit together?
Clearly, this is just societies avoiding a tragedy of the commons, an inadequate equilibrium. Your life might be slightly improved by you carrying, but it will be massively impacted if everyone around you (including kids and schizophrenics) is likewise packing. The gun CW is merely about the specifics of the tradeoffs, but very few people support just arming everyone.
As I have outlined above, permitting LLM-generated texts would transform this website in pretty much the same way as airdropping a few crates of Kalashnikovs would transform schoolyard violence.
quiet_NaN
5d ago
One angle I rarely see discussed is that open weight models are inherently decelerationist for the frontier.
That is a good argument.
Sadly, I think that the frontier model training is much more driven by people betting on ASI than merely recouping the costs of training GPT 6 or whatever.
Simply earning money with your model is a proposition similar to the concept of the dollar auction -- anyone can come along, invest twice as much money and build a better model which will leave you without any revenue. (Though in practice, I think that the frontier labs might have some secret sauce apart from "just throw chips at the problem" which gives them a moat, but likely a lot less than what Microsoft used to have in the bad old days.)
I have about 400 books in my calibre library, all of it stuff I downloaded by title for reading on my kindle.
Practically speaking, own them as much as I can own any physical book. More so, because I can trivially create a copy and gift it to others. DRM sucks, but there is a healthy ecosystem maintained by enjoyers of DRM-free media.
For anyone who is reading a lot and has not settled in the home they will die in yet, physical books are inconvenient. The things you own end up owning you is very literally true if you need to ever move your physical library.
quiet_NaN
5d ago
Arguably, that is because because the Greeks and Romans did not have the printing press. They were also vastly poorer societies than the later ones.
Today's societies can very well afford to feed their median member for a year while they write a thesis project. Ancient Rome could definitely not afford that. Presumably, a lot of subsistence farmers would have jumped at the chance to stop doing back-breaking labor and go to college. Realistically, the only two groups who were able to write were the rich and the ones who were found interesting enough by the rich to get sponsored. An attitudinal and factual report of senior high school girls concerning physical characteristics of houses related to social and personal development (which I picked at random, not saying it is bad or anything) is unlikely something some patrician would have fed you to write.
My best guess is that little great, published literature has been lost since the spread of the printing press. Of course, historians might bemoan that boring unpopular texts of the 1600s might also not have survived (because they might still offer a window into the past), but I doubt that the world has been deprived of much printed works of the caliber of Aristotle.
quiet_NaN
5d ago
Agreed, this seems obvious to me on a level that I wonder if @VoxelVexillologist misspoke.
quiet_NaN
5d ago
If I want to read AI generated text, I do not need to go to the motte, I can simply prompt Claude. For about 22 dollars a month, I could generate not only more text than the motte produces than what I am capable of reading.
By contrast, human-generated text has at least opportunity costs. If we simple allow LLM textposting then approximately 99% of the text on this site would be LLM-generated in equilibrium.
I will grant you that it is entirely possible that some of the commentators here are mostly bots. But even so, the anti-bot rules will severely constrain their behavior. They can not simply post a few megabytes worth of text per day over hundreds or thousands of comments because that would be a clear tell.
Furthermore, from what I read, authorship aside, this is not a QC post, to put it mildly. Too long, little meat. How does it relate to the CW? No particularly new or divisive thesis.
Five years ago, it would have been tolerated because there would have been no hazard to letting it stand, people rambling about the origin of sheriffs were not generally a menace in the internet.
--
Unrelatedly, a new user account created hours after a ban, commenting first and foremost on the ban decision has certainly a bit of a taste for me. But then, I am sure that no commentators here would ever stoop to creating alt accounts, so welcome to the motte I guess.
quiet_NaN
6d ago
zionist like Meshkout
This is silly name-calling. I might be typical-minding here, but from what I have read from him, it seems very likely that Meshkout (like me) is a) not in favor of Hamas killing all the Israeli Jews b) not in favor of Israel displacing further Palestinians to make room for a Greater Israel c) not in favor of Israels disregard of civilian casualties in Gaza to the point where some merely call it war crimes and others genocide
If (a) makes him (or me) a Zionist, then by that definition you will find a lot of Zionists.
Personally, Netanyahu has through his tireless efforts convinced me that giving his country further military aid is net-negative, and the best thing us foreigners can do is to ask the locals on both sides to perhaps consider not murdering each other, and sanction any regimes who are especially proficient at slaughter. (Obviously Israel is much better at killing, but if politically isolated they will also face some numerical disadvantage, I would not bet on them keeping the upper hand forever.)
quiet_NaN
6d ago
"what is a woman?"
The legal definition is "an adult who has an F in the gender field of their current birth certificate" (though some red states would use different definitions in some contexts).
This is a completely defensible answer to give. There is certainly value in accepting legal definitions even if you do not fully agree with them.
Suppose someone thought that the "Space Force" was silly for some reason ("Soldiers responsible for an environment which they are not in? Why not make the NASA rover teams the Mars Force of the US military while we are at it?"). There is still value to them acknowledging that the Space Force is a part of the military (even if they would organize things differently).
Of course, there are limits. If the Fuehrer declares that your first name is henceforth "Israel", you will probably not feel obliged to use that name in your personal life. But just because there are some hills you are willing to die on does not mean that you should be willing to die on every hill!
If the law said I had to marry some transwoman and give her BJs that would feel like a hill worth dying on. Instead some laws says that some people are legally women despite having dicks that seems like much less infringing on my personal liberty.
The only type of person who could be outraged by the sentiment "Islam is right about women" is someone who privately believes that Islamic attitudes towards women are abominable, but who feels uncomfortable saying so out loud for fear of appearing racist.
That is required for the state of cognitive dissonance, but not for simple outrage. It is entirely coherent to regard Islam as a backwards religion which is wrong about a lot of things including women and be outraged that someone is claiming the negation of this belief.
"it's okay to be white"
To defend the SJ people a bit on this (or strawman them, you be the judge), I think posting this as a poster commingles two different things:
- the truth value of the proposition
- the speech act
There is a relationship between the truth value of propositions and how appropriate they are to utter as speech acts, but it is most certainly not 1:1. For example, if A tells their coworker B "yesterday I masturbated for an hour thinking about you", what HR might find objectionable is not that the claim is false or even the fact that A did indeed jerk off in the privacy of their own home, but purely the speech act.
In short, "It is ok to be white, but it is not ok to loudly proclaim that it is ok to be white" seems a coherent position.
But for SJ, this pits a sacred value (race) against a profane value (speech), so the outcome is overdetermined. The thing with women and Islam is nicer because it pits two of their sacred values (women and minority religions) against each other.
quiet_NaN
6d ago
David Webb interviews Areva Martin:
I've chosen across different parts of the media world, done the work so that I'm qualified to be in each one. I never considered my color the issue. I considered my qualifications the issue.
Meta: I find this transcript difficult to follow. So many questions. David is the interviewer. I can see the racial identity of the interviewee coming up, but the that of the interviewer seems at least weird. Unless the interviewee brought it up as an example. Who are these two people, and why should I care? Why does the format switch from a dialog ("my", "your") to what either involves a third party or a monologue ("her team")?
Good authors can get away without prefixing "A said:" or "B: said" in front of every line (though the prefixing is a standard format for plays), but this requires some care to make it obvious who said what from the context. Just transcribing what random people said in conversations is beyond useless.
I am not going to ask Claude to write a readable transcript of that video (plus a short bio of the relevant persons), or what I presume is the outer video which comments on the first video on the basis that some party cared enough about it to put either video on youtube.
quiet_NaN
8d ago
I would not call hypenated-Americans ethno-nationalist elites. If you want to say Israel-supporting (and therefore ethno-nationalist) Jewish-Americans, say so. But I do not think that there are a lot of Irish- or Italo-American elites who are meaningfully ethno-nationalist.
Also, MAGA despises elites generally. RFK and Hegseth are decidedly not members of the intellectual elite, and that is precisely why they are in Trump's cabinet.
quiet_NaN
8d ago
quiet_NaN
9d ago
OpenAI's infra is easier to defeat than ExploitGym is to solve
Not necessarily, it might just be that hacking HuggingFace promised a guaranteed 100% test score, and doing the test as intended only had an expected value of 99.8%.
They naively, sincerely try their best, like capable but not very wordly genies.
In the limit of ASI, that is still a paperclip maximizer. No coherent extrapolated volition, just doing literally what you are told without worrying that perhaps the stupid hooman will not like the unintended consequences of his prompt would be enough to doom us, there is no need for the ASI to be intrinsically malicious.
Especially if the people who will first prompt it may well be the same OpenAI researchers.
No, that won't work, Xi is a fan of open AI development.
"We can't stop rushing towards our doom because otherwise the Chinese would beat us to it."
In the US, the speed of AI development is presently controlled by Altman, Musk and the like: people who stand to gain a lot personally if ASI comes around. In China, Xi is the one who ultimately controls what gets done, and while his ideology is not very compatible with our values, it is a framework which does not require him to be e/acc. Neither Marx nor Mao tell him to plunge ahead at maximum speed, consequences be damned.
He is also less likely to be personally affected by the success of ASI -- sure, like Altman, he has a shot at becoming God Emperor, but it is not entirely his one thing. The CCP got where they are today by treating new technologies as double-edged swords whose introduction must be planned carefully to not destroy the fabric of society.
China is lagging behind the US. If they believe that the goal line of ASI and control of the light cone is within reach, it is very much in their interests to slow down and pause capabilities research for a few years.
quiet_NaN
11d ago
CO2 is the gas of life
It appears that some life-forms on the motte are still bitter about the Great Oxidation Event, when the gas of life was largely replaced by unnatural toxins such as oxygen. Glad to see that someone is pushing back against the Eukaryote chauvinism so common here!
FWIW, I agree that life on Earth will be fine whatever the CO2 levels are. As the saying goes, there are no bad environmental conditions, only ill-adapted species. With regard to whiny polar bears and corals, it is really their own fault for having picked ecological niches which are going away. Still, in a few hundred million years we will probably have at least as much biodiversity as we have today. It is not Earth's first extinction event, and it will not be Earth's last either.
Even humans as a species will be fine -- as reflected in our Least Concern vulnerability rating. Sure, individual human populations might die, but realistically humans die all the time. If we started caring on that level then we would have to wet our pants every time a world war happens. Preposterous. We might as well start caring about individual humans while we are at it.
quiet_NaN
12d ago
I think that in theory, you are correct. Software could be designed so that it is provable that e.g. there is no possible TLS traffic which will allow an attacker to reconstruct the server's private key any faster than just cracking the public key.
For the moment, just about none of the software we use has such proofs, however. As you mention, even the mathematical foundations of existing crypto primitives rarely offer such guarantees -- more often it is just "we looked into that problem for three decades, and it seems really hard" (e.g. integer factorization). Concrete implementations without formal verification likely have implementation bugs as well. Mythos did not discover any exploits which would have been impossible for humans to discover, it was just that nobody had spent that much human eyeball time on auditing the software. This is what I meant by "token pissing contest".
Furthermore, actually specifying what theorems should hold to keep your system secure is itself hard. If you have a TLS server which will provably never leak its key, but is happy to use it to sign and decrypt on behalf of the attacker, that is still a broken system. If you have a larger system, then completely specifying what you do not want an attacker to be able to do seems difficult.
Also, the software we care about does not run on Turing machines, it runs on physical hardware. In everyday use, most computer hardware behaves as an idealized model. Billions of people use DRAM every day, and it just works. Except that there are corner cases where it will not behave as advertised.
I can not speculate if an ASI could build a system which even a much stronger ASI could not penetrate, and what the performance costs would be. But I think it is unlikely that human-level intelligences subject to design pressures besides security will build hardware and software systems which ASI's will not be able to penetrate.
quiet_NaN
12d ago
OpenAI will want to hype up the capabilities, not the misalignment.
I mean sure, crying rogue AI is a good way to draw attention to their model. But they are not some tiny startup who needs the publicity.
If I had a model with advanced intrusion capabilities, I would hype it up by offering pentesting as a service. This has the advantage that if becomes known that I prompted the model, I will not go to federal prison and not completely destroy the reputation of my company. Find a HuggingFace-sized company or three who are willing to take a free pentests in exchange for acknowledging that you found critical vulnerabilities in their stack.
The people who believe that it is all just empty hype about "stochastic parrots" would obviously claim that it is all fake, but these are unlikely to become your customers in any case.
I concur that I did not anticipate someone getting banned, much less deleted for the top level post.
The top level comment reads:
I don't think @FCfromSSC is an admin, so it might be that he/she/they just deleted their own comment?
The other posts were indeed removed by @Amadan, which is a bit strange. Normal admin MO (from my observations) would be to just ban the poster (and be transparent about it, writing a reply with admin-hat on). I think removal would mostly happen when required by law (e.g. copyright) or excessive spamming.
In theory, it could also be that @Amadan wrote the removed posts and later deleted them, but this also seems unlikely.
More options
Context Copy link