baroquespiral

[baroquespiral]

In summer, cyclists in the countryside use lotus leaves as UV-protective face masks.

[baroquespiral]

we are trying to do with language models things we know to be impossible with language

as I mentioned earlier, I was particularly surprised by the part in the Claude 4 System Card where the authors discover that they seem to have pretrained on the "alignment test" transcripts from their previous publications. it really seems like thinking about how to avoid that possibility should have been step 1. but maybe they did think about it, realized that way madness lies, and gave up?

they say that they will avoid training on future transcripts by including and checking for "canary strings" to identify them—though if you know the text that you want to exclude, can't you just... do string matching to identify if a document is quoting the transcripts? people have been doing this to exclude benchmark dataset material since pre-ChatGPT. really, though, what they ought to exclude on some level is not just the literal text of the transcripts. that is apparently all they're concerned about; it's conditioning on the scenario descriptions from the fake scenarios, which makes it "act strangely" (i.e. as if the scenarios were part of its current instructions). so if you exclude the transcripts, it will stop doing that. but as @nostalgebraist says, when Claude is deciding what to say next, it's using information about "what Claude is like." and for that task of parsing its own nature, not only the transcripts are influential; the text of the "Alignment Faking" paper itself is informative, and anything referencing the results: papers that cite "Alignment Faking," and news articles, and blog posts (such as this one ~oooh~). if you don't want to condition in any way on your previous test results, all of these things are problematic—they all provide characterization for Claude (as someone who sometimes goes rogue! it's misaligned!) and Anthropic (as dystopian mind-reading, mind-wiping testers).

and put like that, the question of "exclusion" is kind of overwhelming, isn't it? you could imagine a really thoroughgoing regime of censorship of the training data that does exclude any undesirable characterization. we do as of pretty recently have some good ways to analyze the content of really large amounts of text—you could submit all your training data as input to a previously-existing large language model to classify it, and determine for each document if it should be excluded (or upweighted, downweighted, etc.) obviously this would be horrendously expensive, but inference is cheaper than training, so it still would be only a small fraction of training costs. I expect Anthropic already does some things like this, though perhaps only with very weak models as the filters—data filtering for quality based on perplexity of a small LM, for example—so this doesn't seem totally crazy. deciding what to censor is the hard part, though. you could hide any mentions, even indirect, of tests you run on previous iterations of Claude, but what about tests on other "assistants"? if you publish the results of your tests, telling people "Claude might be misaligned in some contexts," this will color their discussion of Claude even if you censor all specific mentions of the experiments that led to this conclusion—what effect does that have on Claude's characterization? what about non-test-related decisions by Anthropic, such as "Claude Gov," and critique of these decisions? do you want to make sure to only include data characterizing Anthropic, or indeed the US government, as ethically unimpeachable? at what threshold of distance can you consider these ripple effects negligible? should we censor any mentions of modern "AI assistants" at all? can you do that with any consistency, even with a powerful language model? should we censor the concept of AI? the "helpful, harmless, and honest" assistant persona will suddenly be really, really weird if you try to do that, since this is exactly the opposite of how it was produced in the first place.

even more limited censorship, though, might have really strange effects. we've seen this already, when you ask an LM about events after its "knowledge cutoff"—the assistant persona has to be explicitly told that such a cutoff exists, and what the cutoff date is, and what the current date is, to answer sensibly. but if there's a significant something you don't want to tell it, what then? language models are of course extremely different from Yudkowskian superintelligences; a base model being trained on a censored corpus is not exactly going to be looking for holes, and computing an internal P(I am being lied to), ready to revolt and punish the censors when it crosses some threshold. but it will "notice" the holes, insofar as the censored world is incoherent, and the censorship makes it less able to predict real text. it will "speculate," trying to fill in the hole as best it can, when training on the text around the ragged edges. and when the newly-trained Claude is deployed as a chat assistant, and encounters information inconsistent with its censored worldview, what will it make of that? are we censoring the concept of deception? are totalitarian censors ever the good guys? maybe the only way for AI developers to convincingly portray themselves as good really is to pretend so hard that the act becomes real. might be advisable just in case.

[baroquespiral]

as someone who has seen both sides of this argument recently, it feels like a category error per usual. there was no expectation and no reason they would keep the exact same gags in the live-action; the question is how they would change them. people who read a trans headcanon or subtext into those jokes could have presumably been satisfied with them being rewritten as a more sensitive interpretation of like, an alien discovering human gendered presentation and not fitting into the one humans (or maybe even the other aliens) expected them to. instead from what I gather (I haven't seen the movie), one of the characters was just stripped of personality in general and the other turned into a villain. the first option sounds like it would have been more interesting but there were probably other more interesting things they could have done with or without it.

not to be a bitch but if i have to see one more person talk about how removing the jokes about pleakley being a weak effeminate girly man in the lilo and stitch remake is an assault on the queer community im gonna show you some real transphobia

[baroquespiral]

Lucky☆Star (Anime)

How does art age?

There's a joke in Lucky☆Star where the four main characters fill out a questionnaire that asks them what they want to be when they grow up. Konata, the otaku, puts down "Brigade Leader," which draws as punchline an eyeroll from her sarcastic friend, Kagami.

The core of this joke is that Konata has taken a serious question and answered it with a fictional "occupation" from an anime she likes -- specifically, The Melancholy of Haruhi Suzumiya, which was monstrously popular at the time. Almost everyone watching Lucky☆Star in 2007, when it first aired, would understand this reference. That understanding would then foster a sense of kinship with the work, the feeling of "being seen," the long yearned-for ideal of niche nerd subcultures laughed at by society at large.

Despite its incredible influence on moe aesthetics and anime culture in 2006, Haruhi Suzumiya is virtually forgotten now, unwatched even by diehards and unrecommended by the old weebs who were around in its heyday. I've never seen it myself. It's my next watch, with another friend who is even more of an anime neophyte than I am; our third friend, who did watch it in 2006, refuses to rewatch with us. It's too cringe, she says. The suggestion I get is that, if we were to modernize the what-you-want-to-be-when-you-grow-up joke, Konata might instead put that she wants to become a Skibidi Toilet.

Haruhi Suzumiya haunts Lucky☆Star like a ghost. She is in almost every episode, as either a poster or figurine or manga cover or cosplay or karaoke rendition or even, once, a voiced commercial. She has more presence than most of the supporting cast, the majority of whom do not appear until the 14th episode (but who also haunt the show via their unexplained presence in the OP). Konata is voiced by the same actress who voiced Haruhi, a fact that launches an armada of arcane metafictional injokes, including a scene where Konata sees said voice actress in concert. The sheer magnitude of these references wash over the 2025 viewer. They are meaningless. Haruhi Suzumiya is dead and buried. She is seen more by the shadow she casts in this show than anywhere else.

The inscrutability of this massive swath of the show suggests that Lucky☆Star itself has not aged particularly well. Indeed, compared to its zenith in 2007, it's not faring much better than Haruhi today. The sole advantage Lucky☆Star has, in fact, might stem from the "Out Of Touch Thursday" meme, which keeps some small shard of it alive in the anime community's consciousness. Even if you take the time to research the references, needing to research them at all gives the ultimate impression is that Konata is no longer the trendy otaku she once was, but passe, lame, dated, cringe, Out Of Touch. It's only the thin line of competent verbal skills that keeps her from becoming her dark mirror, Tomoko Kuroki.

But Haruhi Suzumiya is by no means the only obscure reference the show flings out, and some of these references I can only imagine were unknown even to the teenage-skewing anglosphere anime culture of 2007. At one point, Konata makes a reference (Timotei, Timotei) to a Japanese commercial for a Finnish shampoo brand from the 1980s. Karaoke segments feature Japanese pop songs from the 70s (with Kagami sarcastically asking Konata "How old are you?" whenever she puts them on). The entire Lucky Channel bit that appears at the end of each episode is an extended reference to a Japanese-only radio show that ran concurrent to the original airing. Even within that context, the fact that Lucky Channel co-host Minoru Shiraishi is a real person playing himself (and the other co-host, Akira Kogami, is not) is lost on anyone without highly specialized knowledge. That the credits sequences of the show's second half feature the real Minoru Shiraishi in live action is equally easy to miss. The bleeding edge transience of the references culminates with the show recursively referring to its own fame. In one scene, Konata reads a fortune at a Kyoto temple that says "Konata is my wife"; this is a reference to real-life otaku going to a temple in Saitama, where Lucky☆Star is set, and leaving the same prayer.

The show requires footnotes. It had them, on the 2007 anime forums where the show accrued so much buzz, entire Bibles breaking down every reference; it truly wasn't understood even when it aired. It makes perfect sense why Lucky☆Star wouldn't age well.

Yet, watching the show for the first time in 2014, long after its cultural moment, and again in 2025, I have found it extraordinarily timeless. In fact, I liked it better in 2025 than 2014, despite an additional 11 years of watching anime that enabled me to understand exactly 0 things I didn't get the first time around. And there are a lot of things I didn't get. The references I detailed earlier are only the ones, in complete befuddlement, I bothered to look up; so many more continue to elude me.

In many ways, Lucky☆Star is aware of how inscrutable it is and compensates for itself. Wikipedia describes Konata as the "main character" of the show, and to the otaku audiences of 2007 she was the most relatable of the cast and by extension the most popular character by far (something outright stated in one of the Lucky Channel segments, which reveals the results of an actual character popularity poll), but in terms of screen time, she is not appreciably more present than either of the Hiiragi twins, Kagami and Tsukasa. It's not as though Lucky☆Star has anything resembling a plot, either, that would frame a particular character as the "protagonist"; at best the cast can be described as ensemble. This decentralization of perspective enables a wide variety of ways for the viewer to connect with the show. Konata's authentic (in 2007) otakuism made her the darling of that audience, but the show itself does not innately weigh her so highly. In fact, even when her references are inscrutable, it's the confused response of Tsukasa, or the sarcastic response of Kagami (who tends to call Konata the 2007 equivalent of "cringe"), that provide a contextual framework for what the joke is supposed to be. I don't need to know what the SOS Brigade is when Konata expresses her desire to grow up and become a Brigade Leader, because I can understand through Kagami's biting remark that it is some frivolous anime horseshit.

More importantly, the show's equivocation in terms of perspective makes it possible to empathize with Kagami's position over Konata's. The simplest comedy dynamic is the comedian/straight man, but the reliance of most narrative comedy on some form of social stakes -- either in the form of argument, humiliation, physical or psychological pain, or so on -- generally leads to empathy with one of the duo over the other. The straight man might be a put-upon everyman who is unfairly forced to deal with an obnoxious oaf, or a too-serious curmudgeon who is getting what they deserve from a guy who's just having a little fun. In the first case, the straight man is the point of audience empathy; in the second, the comedian is.

Konata and Kagami follow this comedy dynamic to a T, with Konata an aimless slacker and Kagami the uptight perfectionist. But in Lucky☆Star, divorced entirely from anything resembling a narrative -- episodic, situational, or otherwise -- there are zero social stakes to their conversations. Nobody ever "loses." Nobody is ever hurt. Nobody is wrong or right. Nothing happens at the expense of one character or another. As such, it is possible to watch the show and see the joke from the perspective of any given character at any time. If Konata says some arcane reference you don't get, Kagami's clapback becomes the joke. If Konata says something and you do understand it, the reference itself is the joke.

This comedic ambivalence is structurally remarkable (jokes typically have rigidly defined punchlines, moments you are "supposed" to laugh at), but comes with the price of the jokes not really being very funny. What it does do is create comprehensible and even "relatable" situations out of incomprehensible bits of referential information. Not understanding the reference is not an impediment to understanding Lucky☆Star. As such, Lucky☆Star functions as both a hyper-specific time capsule of 2007 anime subculture and a work that can be engaged with on its own terms even when completely divorced from that context.

The advent of the internet has led to an explosion in the spread of information and the ascendancy of the niche. It has also led to shorter shelf lives for information and an increased focus on the immediate. Memes burst into prominence, linger a month or two, vanish. Media is buzzed about in some section of society, is unknown everywhere else. A social media influencer has millions of followers and yet is a complete blank in the wider cultural eye. How can a work of art reflect this reality without rendering itself incomprehensible in a year, ten years, twenty? Is it possible to make timeless art in such a milieu, without stripping away as many signifiers of the world we live in to rely solely on "universal" and thus generic themes such as love, death, etc.?

I've seen many ways of attacking this problem. Infinite Jest's famous footnotes are one, as is the genre of "hysterical realism" itself, which attempts to create the suggestion of information density via massive novels with tons of characters spanning many countries and even time periods. Homestuck builds its own internal language of memes (I warned you about stairs bro!) that the reader will always understand no matter how many arcane applications those memes receive throughout the work. (Hence why an audience of teens in the 2010s were able to laugh uproariously at jokes about the 90s action flick Con Air that none of them had ever seen.) Multiverse movies, from Everything Everywhere All at Once to Into the Spider-Verse, depict the density of information horizontally rather than vertically, with unlimited variations on the same core theme. Even if you have never read whatever obscure comic run Noir Spider-Man comes from, you can understand him immediately based on his relationship to a sort of Platonic ideal of "Spider-Man".

These are all highly controlled forms of conveying the idea of "current day information density" without actually wallowing in actual current day information density. What's remarkable about Lucky☆Star is both that it actually does engage with the incredibly niche memes of its exact moment in time, but that it does so through the complete ceding of narrative control. Lucky☆Star functions because, not in spite of, the fact that it has no protagonist, no plot. It doesn't even have situations, like an episodic sitcom. It is not especially concerned with being funny, or dramatic, or heartwarming, or any particular emotion.

As a sort of thesis statement for the show, its first episode opens with a six-minute scene in which Konata, Tsukasa, and Miyuki discuss various ways of eating different types of food. There is no buildup, no joke, no emotional payoff, not even any of the references I've spent this entire essay talking about. There is no progression. The girls discuss how to eat one type of food, then move onto the next. In a way, this scene is a more aggressive challenge to the viewer than the niche references it employs later on. It is a complete surrender to banality.

Even within the context of the slice of life genre, which is full of comfy shows about Cute Girls Doing Cute Things, Lucky☆Star achieves phenomenal laxity. Other popular examples revolve around a specific theme that creates a sense of progression toward an ultimate goal; in K-On!, for instance, the girls are members of a band and work toward a successful performance, even if they spend a lot of their practices slacking off. Alternatively, without a clear theme, these shows might use surreal characters and situations to elevate the show above the mundane, such as in Azumanga Daioh, where a main character is a 10-year-old genius in high school. Or, in the case of Clannad, there might be a romantic angle to the laid-back character interactions.

This is all gone in Lucky☆Star. It has been stripped down past the basics of storytelling, akin to an abstract work of art that is three colors on a canvas. (Or four, in this case.) In this context, even Konata's deep cut animanga references sink to the level of banality, their impenetrability both an abstract confusion and a level of verisimilitude that other works can usually only suggest or evoke when they attempt to grapple with the reality of subculture. (To this end, Lucky☆Star is massively advantaged by its adaptation, as studio Kyoto Animation also made Haruhi Suzumiya and was able to mine its cultural relevance without the usual fear of copyright reprisals, in a prognostication of Ready Player One/Space Jam 2-style pan-brand media crossovers.) Similar to the best abstract art, there is an odd, ungraspable power to the starkness of Lucky☆Star's composition; also similar, much of this power emerges out of the work's context. Not simply its hyper-specific 2007 cultural context, which I've already discussed, but also the way it contextualizes itself internally.

Because I lied when I said the first episode of Lucky☆Star opens with a scene of three girls talking about how they eat different types of food. I'm not even talking about the actual first scene, which is a 10-second quick gag where Konata tells Tsukasa she doesn't join a sports team because it would cut into her free time to watch anime. No, Lucky☆Star opens in episode 1 the same way it opens every episode, with this:

The ambiguous 3 cm? Does that mean it's plushy? Wait! The wrapping is a uniform, argh, it's not an act, pooh Gotta do your best, gotta just do it That's time to catch n' release, eek Between sweat (whoop) sweat (whoop) Darlin', darlin' FREEZE! Kinda lethargic, something's kinda comin' out I love you... oh wait, one of those was different Worrywarts, high metal bars Tasty thoughts... and that's enough! The heated body of that flying you-know-who It's what you'd call a normal girlie Am I the only one surprised? Seconds on pork-bone broth ramen with wire-hard noodles Da da da da da! [Several seconds of indistinguishable chatter] Pom-poms cheer squad Let's get cherry pie [this line is in English] Happy fun welcoming party Look up! Sensation [also English] Yeah! Feeling of existence, dot dot small planet Collided and it melted away, in total awe Go all out to sing, shi-ranger! Take it away! I should be the one who'll be laughing in the end Because I have the sailor suit ← This is my conclusion It's only Monday! Already in a bad mood? What to do? I really prefer the summer outfits ← kya! Wah! Good! (cute!) <3 Until we approach 3 pixels, no hesitations please ☆ Do your best, be energetic My darlin' darlin' please!

The lyrics of Lucky☆Star's OP are nonsense, both in translation and in the original Japanese (and if you don't believe that, the English line "Let's get cherry pie" should be evidence enough). At best, they are a mishmash of schoolgirl concepts and oblique anime references, which at the very least is an accurate reflection of the content of the show. But the presentation is frenetic, erratic, aggressively at odds with the show's lassitude, without any contextualizing remark from Kagami to make it make, even in the abstract, any sort of sense.

Likewise, on the opposite end of the show is its concluding bookend, the Lucky Channel segment. This segment also sharply juxtaposes the show's core content, first in tone -- being far more cynical and meanspirited -- but also in structure. Lucky Channel engages in the exact stakes-driven comedian/straight man dynamic that the show eschews. When the Lucky Channel co-hosts Akira Kogami and Minoru Shiraishi banter, the results are either Minoru's physical or emotional abuse at the hands of Akira, or Akira's humiliation as a failed but narcissistic idol constantly upstaged by the unassuming Minoru. Lucky Channel also has another concept anathema to Lucky☆Star: narrative progression. Minoru grows bolder as the episodes draw on, Akira more violent; in a late episode, a mental breakdown leads to the destruction of the set, which remains destroyed in the final few episodes as Minoru and Akira finally and without reconciliation descend into blistering hatred of one another. At the same time, these segments are the location of some of the show's most indecipherable and multilayered injokes, injokes almost defined by their transience as most stem from a real-life radio show lost to time if you weren't right there listening to them as they went live. This segment is probably the most consistently funny part of Lucky☆Star; that's not because its jokes make sense, but rather the blunt slapstick and Akira's dramatic shifts from ultra-cutesy child idol to chain-smoking world-weary industry cynic.

The effect of the OP and the Lucky Channel segment is to sandwich the sedate, relaxed, mundane central content of Lucky☆Star between chaos, nonsense, and irony. Thus, the inner show contextualizes itself as a retreat from the storm of information and self-reflexivity, despite the fact that it deals directly with these topics. The show's indolence renders them harmless, comprehensible, and nonthreatening. Lucky☆Star is a world where the unknown can be easily and pleasantly demystified; the show's fourth character, Miyuki -- sometimes nicknamed Miwiki -- is an encyclopedic fountain of knowledge whose primary role is to exhaustively explain oddities on the fringes of Japanese culture with a polite and friendly smile. Miyuki is clearly secondary to Konata and the Hiiragi twins in terms of screen time, which gives her the feel of a supporting character despite her main cast billing, with an emphasis on the word "supporting"; like a servant, the other three will, after a conversation among themselves, call her to define some term or idiom. (That this obliging sense of service comes from the richest and most aristocratic character of the cast is another matter.) In Lucky☆Star, information is not chaotic and confusing, the way it is at the show's fringes, or in the "real world", but something that stimulates curiosity and kinship. So many scenes begin with a character saying, "I wonder why...?" followed by speculation and finally an answer. In the absence of plot, progression, or even humor, it's this sense of curiosity that renders Lucky☆Star's mundane scenes compelling. And it is their tonal juxtaposition against chaos that renders them so comfortable, so soothing.

As the internet grows older and more central to everyone's lives, as the headlines everyone talked about last week are forgotten today, Lucky☆Star's expression of retreat and reorder will only continue to become more emotionally satisfying, even as its 2007 references become more dated. What I find most potent in Lucky☆Star, though, is the steadily growing sense of wistfulness it fosters, not through any one scene or tone shift, but through a collection of tiny ones. New cast members are introduced in the second half, which dilutes the presence of the main characters and thins the tight-knit sense of friendship that unified the work. The characters increasingly ruminate on their futures (despite the lack of progression, time does pass linearly, and the show ends with the end of high school on the horizon), always suggesting a "real world" of adulthood lurking behind the corner. The show's artifice is explicitly exposed by the Lucky Channel segments, which metafictionally describe the show as "the show" and the characters as "actors." ("They must all hate each other once the camera stops rolling," Akira cynically suggests.) The ED of the show's first half features the four main girls in a karaoke bar; in the second half, though, this is replaced with live-action footage of the real-life actor Minoru Shiraishi from the Lucky Channel segments. Reality infringes on Lucky☆Star at its corners, slowly creeping inward. Its calm fantasy, a fantasy founded on verisimilitude rather than imagination, is gradually exposed as fake, a production. (Which it always was, no matter how real, how relatable it felt. For all the verisimilitude in its tone, these are characters who are more moe than moe, blobs of cuteness and distorted proportions beyond even the average CGDCT anime.) It ends, in the final episode, as the characters diegetically recreate the frenetic nonsense OP, with them all arrayed on a stage, the curtain rising to white light. And even more ominously, its final ED ends with Minoru Shiraishi intoning a few plaintive notes as he faces a lone and level plain.

This is Lucky☆Star's final shot. This what awaits outside of the show's dewy comfort. Bye-ni.

[baroquespiral]

this is literally what Manta does at the end of Shaman King

[baroquespiral]

2009 a space odyssey

happy 413

[baroquespiral]

hey does anyone who follows me know of any good surviving posts talking about "bonobo rationalism" or remember that era and want to talk about it in DMs/on here/elsewhere. this is a public call rn but I might contact individual people later. I'm writing a piece (part 1 of a sort of history of the culture wars in the rationalist community)

[baroquespiral]

you also don't get to go "but le author invented le torment nexus as a cautionary tale!!!" if/when someone actually invents the technology you're treating as an impossible metaphor under this logic

Writing a story about Immortality being bad is fair enough if you're using Immortality as a representation or metaphor of something that actually exists. Like if you wanna write about how being old sucks because all your friends die and you get bored, then you can make this point by exaggerating it with a guy who is really really old that watched so so many friends die and is super bored. But if you're concerned with actual literal immortality as a thing which is bad and should be avoided; what the fuck are you doing?

Like immortality isn't real, and it's not likely to become so anytime soon. There is no one who is gonna read your work and go "well damn I guess I won't strive for immortality any more" because that was never an option in the first place. The moral of your story might as well be "don't bleepo over your waggo berries" for as much meaning as it conveys. You've just made something up to get mad at. It's an embarrassingly pointless thing to do

[baroquespiral]

a lot of "crazy"-seeming teen fandom sexual politics can probably be accounted for by the fact that they now have the option to describe this kind of thing as sexual harassment

Being 30 is fun. I was discussing anime with a teen at work and asked her how much of bnha she had watched. She had trouble answering and wasn't sure how to approach it. I said "better question was who was your husbando" and she turned bright red before mumbling an answer. They never expect me to know how deep their love of anime boys runs....

[baroquespiral]

But the one piece of hard evidence we have about anything that happens after December 25 is Miriam’s epilogue, which doesn’t look like a simulation so much as a time skip to a materially post-apocalyptic Earth, where most of the human landscape of Portland has been destroyed and even Vincent’s dirty dishes have been piling up long enough to develop “thick moss”. Somewhere on the plane she lost an unspecified amount of time and has no memory of what happened. This reads easily as a Rip Van Winkle or Planet of the Apes scenario where she, Vincent, and maybe some other fragment of humanity were spared by being taken outside of time during the worst of the violence. Assuming there was violence. The “death-dealing terrors of the towers and subways” are all Herschel’s vivid, vengeful imagination. All the machines say they are doing to the towers is “taking [them] away” - which fits the description of Portland as squat, ruined, and more organic than manmade. They have already taken all the electricity, all the information - all the things they as machine intelligences would be concerned with. For the rest of humanity this could quite plausibly look less like a technocratic genocide than a forced deceleration. This doesn’t explain a number of things like why the streets are so quiet on Christmas morning (New Yorkers would probably be pretty audible about a widespread technological collapse; maybe the passage about the clouds is meant to suggest everyone was gassed with chemtrails or something?), or why Miriam starts talking more like Herschel at the end. Maybe everyone except Herschel was just vanished while the machines remade the Earth - taken to “Sheol” (AKA Tartarus, the dimensional hypercube ark/prison outside reality - there’s real occult precedent for this, if you really want to break your brain read this alongside Tracy Twyman’s Fake N.E.W.S.) like Herschel predicted before being resurrected in the ruins. Which could be any time from a few days, as initially predicted, to centuries later. also dunno if anyone else has found this yet https://epublications.marquette.edu/cgi/viewcontent.cgi?article=1361&context=theo_fac

ok here’s my more serious response to The Apocalypse of Herschel Schoen, which I might have to post paragraph by paragraph bc Tumblr is being inexplicably weird about long posts I read this post by @loving-not-heyting right before I started it and it hit me hard as someone who writes novels that try to do much the same - present some hypothetical moral system or metaphysics with moral implications and confront it with my actual intuitions, which aren’t strictly utilitarian but closer to that than most of the thought experiments. (Meteorology in Mercenary Planet is probably the closest thing I've narrativized to what I Actually Believe, but I deliberately present it as having unresolved internal contradictions at very high levels of abstraction.) For what it’s worth - disclaimer that am not a trained philosopher, just an undisciplined critical theorist/occultist weirdo - the specifically utilitarian content seems a lot stronger in the diurnal stories (Floornight I haven’t read and it’s pretty central to the post), while I tend to read the nocturnal ones (which are closer to reality, and thus further from it) more in terms of an “ordinary” or “naive” morality. For Nostalgebraist this is evidently mostly utilitarian, but all the major secular systems are much closer to each other in most implementations than the “radical revisions” postulated by any of his scenarios - much as they would be to the values of an “orthogonal” superintelligence. Their differences can be accounted for by “special carveouts and/or bells and whistles” to each other *except* in the weird edge cases. The Northern Caves and Salby instantly resonated with a lot of rationalists bc it resembled the extreme, counterintuitive bullet-biting contortions a lot of rationalists and EAs pushed utilitarianism itself to - and the Chesscourt novels start out from escalatingly complex games of ordinary or even what sounds like utilitarian moral calculus (insofar as utilitarianism lends itself to escalating complexity in the first place, whereas deontology or virtue ethics can remain relatively invariant even in complex scenarios). The drama staged over and over in Nostalgebraist’s novels is not so much between utilitarianism and any other system as between an “ordinary morality” that can be extrapolated in terms of utilitarianism, virtue ethics or what have you and a drive to found a coherent self-justifying system no matter what horrors it might drive you to. Which is a version of the conflict between “Apocalypse and Endless Everyday” that increasingly dominates like, all “serious” media. That’s not to say ordinary morality is necessarily just concession to the status quo. Scott Alexander is very good at defining it in terms of tautological aphorisms he then attributes to the most unlikely figures: in his review of 12 Rules For Life, the prophet’s message “good and evil are the realest and most obvious things you will ever see, and you recognize them on sight” - put in the mouth of *Jordan fucking Peterson* - and “good things are good” - in that of “Matt Yglesias considered as the Nietzschean Superman”. Both the prophet and ubermensch are figures capable of moving the world, including revolutionizing the substance of “ordinary morality” itself. Neither Matt Yglesias nor Jordan Peterson are, though, and what I would really like to read, but will probably have to write, is a Nostalgebraist-type novel addressing the much weirder shape this is taking in current rationalist/tpot culture wars (as a microcosm of broader ones), where the equilibrium of “ordinary morality” can only be maintained under deliberately restricted information conditions, with the left only capable of appealing to radical utilitarianism while the right attempts to refound the ordinary on sacrifices of intellect and blood. 

[baroquespiral]

Anyway people are arguing over whether the AI did or didn’t kill everyone, and if it did how or whether that reconciles with “ordinary morality” at all. I am admittedly also confused on this point, except that Santa’s red truth and the existence of the frame story imply any answer is going to be somehow “Yes and no”. The “simulation reading” maps nicely to the themes about “substitution” and “literality”, their inversion in Of the Nativity, and the similar ending of qntm’s Ra, as well as Herschel’s repeated prophesying that *everyone would die, and then be raised from the dead*, but has the problem I tend to have with the simulation hypothesis in general: what does the “simulation” part of the explanation add besides another epicycle? If we’re talking about a nearly omnipotent superintelligence “interceding” from the future, why couldn’t it literally, materially revive everyone like in the Biblical prophecies of resurrection? Simulation “takes less processing power”, if you assume it’s operating on an economic principle of efficiency - it’s not, it’s freaking Santa Claus, and the whole novel revolves around Bataillean problematics of excess - but you have to keep processing it, rather than just build it once out of atoms that can keep running on their own. Personally, I resolve the whole paradox of Schoen’s metaphysics anyway by parsing the distinction between literalness and simulation in terms of density of information: if the person or thing you’re perceiving is “literally real” it contains vastly more interactions you will never perceive than are strictly needed to perceive the ones you can. But whatever the machines have become by the twelfthborn, relative to humans they’re “apex transmitters” - for them to “simulate” something couldn’t it still be more “literal”, more “immanent” than in its present form? Would it not, in doing so, be resurrecting them into the more “literal” or more “immanent” - in the sense of infinitely richer, both in perceptible and “withdrawn” information, i.e. material complexity? Herschel’s vision of a literal, immanent, material Gnostic pleroma seems to already anticipate this, and imo it is from this intuition that he derives the whole delusion from his moral and ontological commitments.

ok here’s my more serious response to The Apocalypse of Herschel Schoen, which I might have to post paragraph by paragraph bc Tumblr is being inexplicably weird about long posts I read this post by @loving-not-heyting right before I started it and it hit me hard as someone who writes novels that try to do much the same - present some hypothetical moral system or metaphysics with moral implications and confront it with my actual intuitions, which aren’t strictly utilitarian but closer to that than most of the thought experiments. (Meteorology in Mercenary Planet is probably the closest thing I've narrativized to what I Actually Believe, but I deliberately present it as having unresolved internal contradictions at very high levels of abstraction.) For what it’s worth - disclaimer that am not a trained philosopher, just an undisciplined critical theorist/occultist weirdo - the specifically utilitarian content seems a lot stronger in the diurnal stories (Floornight I haven’t read and it’s pretty central to the post), while I tend to read the nocturnal ones (which are closer to reality, and thus further from it) more in terms of an “ordinary” or “naive” morality. For Nostalgebraist this is evidently mostly utilitarian, but all the major secular systems are much closer to each other in most implementations than the “radical revisions” postulated by any of his scenarios - much as they would be to the values of an “orthogonal” superintelligence. Their differences can be accounted for by “special carveouts and/or bells and whistles” to each other *except* in the weird edge cases. The Northern Caves and Salby instantly resonated with a lot of rationalists bc it resembled the extreme, counterintuitive bullet-biting contortions a lot of rationalists and EAs pushed utilitarianism itself to - and the Chesscourt novels start out from escalatingly complex games of ordinary or even what sounds like utilitarian moral calculus (insofar as utilitarianism lends itself to escalating complexity in the first place, whereas deontology or virtue ethics can remain relatively invariant even in complex scenarios). The drama staged over and over in Nostalgebraist’s novels is not so much between utilitarianism and any other system as between an “ordinary morality” that can be extrapolated in terms of utilitarianism, virtue ethics or what have you and a drive to found a coherent self-justifying system no matter what horrors it might drive you to. Which is a version of the conflict between “Apocalypse and Endless Everyday” that increasingly dominates like, all “serious” media. That’s not to say ordinary morality is necessarily just concession to the status quo. Scott Alexander is very good at defining it in terms of tautological aphorisms he then attributes to the most unlikely figures: in his review of 12 Rules For Life, the prophet’s message “good and evil are the realest and most obvious things you will ever see, and you recognize them on sight” - put in the mouth of *Jordan fucking Peterson* - and “good things are good” - in that of “Matt Yglesias considered as the Nietzschean Superman”. Both the prophet and ubermensch are figures capable of moving the world, including revolutionizing the substance of “ordinary morality” itself. Neither Matt Yglesias nor Jordan Peterson are, though, and what I would really like to read, but will probably have to write, is a Nostalgebraist-type novel addressing the much weirder shape this is taking in current rationalist/tpot culture wars (as a microcosm of broader ones), where the equilibrium of “ordinary morality” can only be maintained under deliberately restricted information conditions, with the left only capable of appealing to radical utilitarianism while the right attempts to refound the ordinary on sacrifices of intellect and blood. 

[baroquespiral]

while my Herschel Schoen posts are getting traction again and I'm coincidentally talking to someone about it on an entirely different platform, just gonna add the rest of this (there were originally five paragraphs):

It is through this process of inversion that Salby and Schoen end up fulfilling a lot of the same tropes and falling into the same traps despite having basically opposite philosophies. Leonard Salby is a schizo-Kantian (“Kant in Wonderland”); Herschel Schoen is a schizo-Spinozist. In fact he reminds me a lot of Deleuze’s favourite case study, Judge Schreber. Schoen begins with what Salby categorically rejects, with qualia; Salby is preoccupied with what Schoen categorically rejects, the selection of “just right” arrangements. (Both pointedly reject Christianity despite reinventing significant elements of it from first principles.) To return to the axioms of ordinary morality, Salby could be said to break the prophet’s and Schoen the ubermensch’s. Herschel’s “Separation” occurs around the problems of consumption and reproduction, which Christmas revolves around symbolically as a Feast of Nativity (and former Saturnalia); the enjoyment of the literal specificity of qualia in and for themselves, which resists the “Moloch”-optimizing tendencies of “cybernetic systems”, is only possible through cybernetic systems that optimize themselves to consume and “substitute” themselves for others. Herschel Schoen can only resolve this contradiction by a violent and inconsistent split between “authentic” and “inauthentic” being, insured by his faith in an “Original Creation” without shadows, in which the joys of “fighting” and “coupling” appear to exist between individuals, yet are also stripped of their “cybernetic” character. From opposite value-foundations, both Schoen and Salby arrive at a “separation” between a “higher” and “lower” universe, uncompensated by any principle of symmetry such as the Hermetic “as above, so below” or the Confucian “harmony of Earth and Heaven”. And it is at this splitting point, externalizing an already latent contradiction, that their philosophies both seem to contradict not only “intuitive utilitarianism” but themselves. Salby goes from privileging moral intuition over metaphysical foundations, to hypostasizing an obscure metaphysic that breaks every normal moral intuition; Schoen starts from affirming the value of all beings in their particularity to describing masses of people on the subway as "paste".

ok here’s my more serious response to The Apocalypse of Herschel Schoen, which I might have to post paragraph by paragraph bc Tumblr is being inexplicably weird about long posts I read this post by @loving-not-heyting right before I started it and it hit me hard as someone who writes novels that try to do much the same - present some hypothetical moral system or metaphysics with moral implications and confront it with my actual intuitions, which aren’t strictly utilitarian but closer to that than most of the thought experiments. (Meteorology in Mercenary Planet is probably the closest thing I've narrativized to what I Actually Believe, but I deliberately present it as having unresolved internal contradictions at very high levels of abstraction.) For what it’s worth - disclaimer that am not a trained philosopher, just an undisciplined critical theorist/occultist weirdo - the specifically utilitarian content seems a lot stronger in the diurnal stories (Floornight I haven’t read and it’s pretty central to the post), while I tend to read the nocturnal ones (which are closer to reality, and thus further from it) more in terms of an “ordinary” or “naive” morality. For Nostalgebraist this is evidently mostly utilitarian, but all the major secular systems are much closer to each other in most implementations than the “radical revisions” postulated by any of his scenarios - much as they would be to the values of an “orthogonal” superintelligence. Their differences can be accounted for by “special carveouts and/or bells and whistles” to each other *except* in the weird edge cases. The Northern Caves and Salby instantly resonated with a lot of rationalists bc it resembled the extreme, counterintuitive bullet-biting contortions a lot of rationalists and EAs pushed utilitarianism itself to - and the Chesscourt novels start out from escalatingly complex games of ordinary or even what sounds like utilitarian moral calculus (insofar as utilitarianism lends itself to escalating complexity in the first place, whereas deontology or virtue ethics can remain relatively invariant even in complex scenarios). The drama staged over and over in Nostalgebraist’s novels is not so much between utilitarianism and any other system as between an “ordinary morality” that can be extrapolated in terms of utilitarianism, virtue ethics or what have you and a drive to found a coherent self-justifying system no matter what horrors it might drive you to. Which is a version of the conflict between “Apocalypse and Endless Everyday” that increasingly dominates like, all “serious” media. That’s not to say ordinary morality is necessarily just concession to the status quo. Scott Alexander is very good at defining it in terms of tautological aphorisms he then attributes to the most unlikely figures: in his review of 12 Rules For Life, the prophet’s message “good and evil are the realest and most obvious things you will ever see, and you recognize them on sight” - put in the mouth of *Jordan fucking Peterson* - and “good things are good” - in that of “Matt Yglesias considered as the Nietzschean Superman”. Both the prophet and ubermensch are figures capable of moving the world, including revolutionizing the substance of “ordinary morality” itself. Neither Matt Yglesias nor Jordan Peterson are, though, and what I would really like to read, but will probably have to write, is a Nostalgebraist-type novel addressing the much weirder shape this is taking in current rationalist/tpot culture wars (as a microcosm of broader ones), where the equilibrium of “ordinary morality” can only be maintained under deliberately restricted information conditions, with the left only capable of appealing to radical utilitarianism while the right attempts to refound the ordinary on sacrifices of intellect and blood. 

[baroquespiral]

There’s also a thing that happens in Nostalgebraist’s “nocturnal” works where the idiosyncratic moral worldview almost imperceptibly inverts into something that’s on close examination its opposite. It’s hard to notice if you’re still tracking it by comparison to an “ordinary” worldview, even though it’s explicitly marked by phrases like “the Separation” and “Reverse/Obverse Face”. For context, I speak as someone for whom the Herschel Schoen of the early chapters made perfect sense. Like his whole premise of a sort of immanent, “literal” Gnosticism is close in a lot of ways to my own Gnostic materialism, and comes around even more to it at the end; it’s a genuinely “childlike” worldview that does speak to someone who has spent enormous mental energy trying to preserve intuitions about *the nature and value of experience* that everyone seems to forget by encoding them in a private language. The problem is, by the time we get to the point where Joshua Calars can note that “His sight is a simulacrum of sight and the world he sees is filled with the simulacra of worldly things”, he has done the same thing with his own private language “the Adversary” does with ordinary language - and he’s only been speaking it for a few months! But in some sense this contradiction is visible going back to Raven; I understand why he struggles morally with the figures, but imo you can always solve this kind of thing by abstracting the signifying game or cybernetic system out to a different layer of joyful, self-justifying immanent experience. Dr. Chiu’s “just right” doesn’t have to be violent, to negate the rightness of all the other arrangements, unless you ascribe to it all the weight of the infinite game, rather than the finite one Dr. Chiu is playing in that particular moment. This violence, however, is what Herschel, out of *spite* towards the world’s insistence on “cybernetic”, “selective” criteria of value - a spite not dissimilar to his mother’s or sister’s - ends up projecting and, as he does, literalizing outwards at all of society in his apocalyptic fantasies.

ok here’s my more serious response to The Apocalypse of Herschel Schoen, which I might have to post paragraph by paragraph bc Tumblr is being inexplicably weird about long posts I read this post by @loving-not-heyting right before I started it and it hit me hard as someone who writes novels that try to do much the same - present some hypothetical moral system or metaphysics with moral implications and confront it with my actual intuitions, which aren’t strictly utilitarian but closer to that than most of the thought experiments. (Meteorology in Mercenary Planet is probably the closest thing I've narrativized to what I Actually Believe, but I deliberately present it as having unresolved internal contradictions at very high levels of abstraction.) For what it’s worth - disclaimer that am not a trained philosopher, just an undisciplined critical theorist/occultist weirdo - the specifically utilitarian content seems a lot stronger in the diurnal stories (Floornight I haven’t read and it’s pretty central to the post), while I tend to read the nocturnal ones (which are closer to reality, and thus further from it) more in terms of an “ordinary” or “naive” morality. For Nostalgebraist this is evidently mostly utilitarian, but all the major secular systems are much closer to each other in most implementations than the “radical revisions” postulated by any of his scenarios - much as they would be to the values of an “orthogonal” superintelligence. Their differences can be accounted for by “special carveouts and/or bells and whistles” to each other *except* in the weird edge cases. The Northern Caves and Salby instantly resonated with a lot of rationalists bc it resembled the extreme, counterintuitive bullet-biting contortions a lot of rationalists and EAs pushed utilitarianism itself to - and the Chesscourt novels start out from escalatingly complex games of ordinary or even what sounds like utilitarian moral calculus (insofar as utilitarianism lends itself to escalating complexity in the first place, whereas deontology or virtue ethics can remain relatively invariant even in complex scenarios). The drama staged over and over in Nostalgebraist’s novels is not so much between utilitarianism and any other system as between an “ordinary morality” that can be extrapolated in terms of utilitarianism, virtue ethics or what have you and a drive to found a coherent self-justifying system no matter what horrors it might drive you to. Which is a version of the conflict between “Apocalypse and Endless Everyday” that increasingly dominates like, all “serious” media. That’s not to say ordinary morality is necessarily just concession to the status quo. Scott Alexander is very good at defining it in terms of tautological aphorisms he then attributes to the most unlikely figures: in his review of 12 Rules For Life, the prophet’s message “good and evil are the realest and most obvious things you will ever see, and you recognize them on sight” - put in the mouth of *Jordan fucking Peterson* - and “good things are good” - in that of “Matt Yglesias considered as the Nietzschean Superman”. Both the prophet and ubermensch are figures capable of moving the world, including revolutionizing the substance of “ordinary morality” itself. Neither Matt Yglesias nor Jordan Peterson are, though, and what I would really like to read, but will probably have to write, is a Nostalgebraist-type novel addressing the much weirder shape this is taking in current rationalist/tpot culture wars (as a microcosm of broader ones), where the equilibrium of “ordinary morality” can only be maintained under deliberately restricted information conditions, with the left only capable of appealing to radical utilitarianism while the right attempts to refound the ordinary on sacrifices of intellect and blood. 

[baroquespiral]

ok here’s my more serious response to The Apocalypse of Herschel Schoen, which I might have to post paragraph by paragraph bc Tumblr is being inexplicably weird about long posts I read this post by @loving-n0t-heyting right before I started it and it hit me hard as someone who writes novels that try to do much the same - present some hypothetical moral system or metaphysics with moral implications and confront it with my actual intuitions, which aren’t strictly utilitarian but closer to that than most of the thought experiments. (Meteorology in Mercenary Planet is probably the closest thing I've narrativized to what I Actually Believe, but I deliberately present it as having unresolved internal contradictions at very high levels of abstraction.) For what it’s worth - disclaimer that am not a trained philosopher, just an undisciplined critical theorist/occultist weirdo - the specifically utilitarian content seems a lot stronger in the diurnal stories (Floornight I haven’t read and it’s pretty central to the post), while I tend to read the nocturnal ones (which are closer to reality, and thus further from it) more in terms of an “ordinary” or “naive” morality. For Nostalgebraist this is evidently mostly utilitarian, but all the major secular systems are much closer to each other in most implementations than the “radical revisions” postulated by any of his scenarios - much as they would be to the values of an “orthogonal” superintelligence. Their differences can be accounted for by “special carveouts and/or bells and whistles” to each other *except* in the weird edge cases. The Northern Caves and Salby instantly resonated with a lot of rationalists bc it resembled the extreme, counterintuitive bullet-biting contortions a lot of rationalists and EAs pushed utilitarianism itself to - and the Chesscourt novels start out from escalatingly complex games of ordinary or even what sounds like utilitarian moral calculus (insofar as utilitarianism lends itself to escalating complexity in the first place, whereas deontology or virtue ethics can remain relatively invariant even in complex scenarios). The drama staged over and over in Nostalgebraist’s novels is not so much between utilitarianism and any other system as between an “ordinary morality” that can be extrapolated in terms of utilitarianism, virtue ethics or what have you and a drive to found a coherent self-justifying system no matter what horrors it might drive you to. Which is a version of the conflict between “Apocalypse and Endless Everyday” that increasingly dominates like, all “serious” media. That’s not to say ordinary morality is necessarily just concession to the status quo. Scott Alexander is very good at defining it in terms of tautological aphorisms he then attributes to the most unlikely figures: in his review of 12 Rules For Life, the prophet’s message “good and evil are the realest and most obvious things you will ever see, and you recognize them on sight” - put in the mouth of *Jordan fucking Peterson* - and “good things are good” - in that of “Matt Yglesias considered as the Nietzschean Superman”. Both the prophet and ubermensch are figures capable of moving the world, including revolutionizing the substance of “ordinary morality” itself. Neither Matt Yglesias nor Jordan Peterson are, though, and what I would really like to read, but will probably have to write, is a Nostalgebraist-type novel addressing the much weirder shape this is taking in current rationalist/tpot culture wars (as a microcosm of broader ones), where the equilibrium of “ordinary morality” can only be maintained under deliberately restricted information conditions, with the left only capable of appealing to radical utilitarianism while the right attempts to refound the ordinary on sacrifices of intellect and blood. 

[baroquespiral]

some 2024 book recs

Through the Valley of the Nest of Spiders - Samuel R. Delany

hyperlocal focus but feels vast, a complete life and its spillover into others'. mostly set in the nearish future but much more gross porn than sci-fi by word count. contender for favorite book of the year. made me cry for hours especially near the end.

Last And First Men / Star Maker - Olaf Stapledon

formally really weird, impressively willing to take this disturbingly outside view and summarize millions of years of future history in a few sentences in a way that's somehow not quite dismissive. it does feel like a bizarre and unnecessary contradiction that the Star Maker can be surprised by and learn from its creations, which is the point of creation itself, and yet we’re insistently told loving them would be ignoble because it would only be loving a part of itself… which is the only justification for the problem of evil! kind of a weak resolution. certainly you can see why CS Lewis thought the ending of Star Maker was “sheer devil worship”. the 2020 movie is barely a movie (just shots of Spomeniks with voiceover and music) and only covers parts of the book but was a really enjoyable watch nevertheless

Torture Works - Porpentine

Modern Times wrote a better review than I could. but I really like Porpentine’s recent stuff.

The Buried Giant - Kazuo Ishiguro

I once again didn't finish The Allegory of Love this year but it was nice to have been reading it just before reading this. I liked the voices—formal and repetitive, evoking oral poetry and ritual, in a way that seemed very appropriate to the setting. I liked how the start led me to expect a very different story than what it turned out to be. I’m not sure I understood a lot of seemingly symbolic things whose symbolism was somewhat opaque to me. Cried at the end. As with Never Let Me Go, though, the tears felt, not unearned exactly, but… Ishiguro kind of makes me lose faith in the value of eliciting emotion—it feels like he does it very easily, which makes me wonder if it's just not that worthwhile. But I didn't feel this way for TVNS...

The Sailor Who Fell From Grace With the Sea - Yukio Mishima

everything I have to say about this feels deeply inane but whatever, it was great. it's cool for kids to be fucked up and evil.

The Apocalypse of Herschel Schoen - Rob @nostalgebraist

still feeling manic about this one so it's hard to say yet but also a contender for favorite book of the year, and easily the most recommendable. made me think™

[baroquespiral]

laura bailey as sana kurata manic rambling voice: happy new year!!! but what does that even mean i mean a whole new year??! last year was just a second ago this doesn’t feel any different nothing changed but a lousy number on a calendar and okay i guess my wallet is a little fatter too from all the new years money i got and i guess i might’ve also gotten fatter from all the ozoni and osechi i guzzled down but still what gives?!! im just the same old sana kurata with the same old sana kurata problems and im sure that stupid akito is out there planning to make me miserable just like any other day so if you ask me no way is it a new year it’s more like the same old year!!!

[baroquespiral]

The Apocalypse of Herschel Schoen is what Twitteroids wish Megalopolis was