gattsuru
1yr ago
I've been skeptical for a while, but to expand the reasoning:
More specifically: I, a gay, centrist Biden voter, am one of the most conservative students at my law school. The Federalist Society here is anemic and widely derided, while there's a dizzying array of progressive organizations. The professors and administrators are, if anything, even more progressive. My school is in no sense an outlier in this regard, nor is this specific to law. The same patterns are overwhelmingly visible in every group of educated, young professionals.
At the trivial level, it's worth spelling out why that is. Conservatives are a minority among the sort of identification you're talking about. But even when 'yuppie' leaned conservative, this dizzing advantage for progressive organizations still existed (it dates back to at least the Eisenhower era!), and the lean of organizations and visible political speakers today is far greater than that of their underlying demographics.
Conservatives and conservative organizations are not just uncommon but destroyed in a wide variety of professional fields, and that's a result of enemy action. Maybe that's well-intended, sometimes, but more often the good intentions or serious objections are a pretext. More often, it's not even that. Simple discrimination is common enough that it doesn't even have to pretend to hide. Demands to fire anyone to the right of the last Democratic President are common, regularly backed up by violent protest. And that persists outside of academia: state bar selections of continuing education credits have gotten hilarious recently, and one of the single most effective members of the SCOTUS bar got booted from his practice, with the threat leveled during and about an appeal. I can go into further detail if @Amadan wants to do the "you are not oppressed" deal, but it's a long list spanning decades, and I don't think you need me drop thirty examples. You have your recent tweet on Sanderson; you don't need me to spell out how suicidal trying to be a mainstream conservative culturati gets.
That doesn't necessarily make you wrong, but it does change any potential solution. A conservative -- or even anything people want to call conservative, with all that implies -- working within the system is inviting a cheesegrater to their tender bits, hopefully figuratively. Any conservative organization trying to work within the system at minimum is subject to being shut down at a moment's notice, if not subject to being hollowed out and worn like a skin suit; any effective capability itself becoming justification for such an attack.
What does change the conclusion is that Doom bit. There is not some deep physical law that educated young professionals are the source of administrative or executive power in this world. They have been favored for the last seventy years because (outside of academic-enforced Curleyism) they were competent, not just in systems that they created, but in their ability to manage and adapt to the world.
The average college graduate today struggles to use a screwdriver, and increasing numbers struggle to write or comprehend an essay; a far broader group have actively rejected even the ideals of meaningful understanding of reality. Teacher's unions have begged and striked to require increasing levels of education that you and I know does absolutely shit for their actual capability, and they're unusual for anyone studying it, rather than it being a problem. If you throw the mandate of heaven in the trash, it ends up in the trash.
That's not necessarily a good thing! Obviously there's the big grifter problem, where once you realize that the TV-show grifter and the PhD are equally unknowledgable about 1800s history, you have the problem of distinguishing what randos do have anything. There's a lot of infrastructure and cash that's hard to replicate outside of academic or industrial settings, and the resulting processes not getting done because those settings are so hostile to you they'd rather burn cash and credibility, and just no one trying, doesn't change much.
More broadly, there are still places that have keep some undercurrent of adherence to actual skill or knowledge that's hard to develop elsewhere, with some interest in actual capability, whether or not they've been skinsuited by politics. There's a far broader scope where the things they teach aren't deep knowledge or skills, but they're the teacher's passwords necessary to get anywhere today. FCFromSSC-style "iterated harm-seeking" is going to be very interesting in the !!bad!! sort of ways, when applied here.
But the resulting answer is going to look very different than Kulak, or than Theilites, for better and for worse.
gattsuru
1yr ago
I'm going to focus on what I think is your real question, the much more circumstantial "Why is Kissinger hated so much more aggressively than other ghouls and swamp creatures like a Donald Rumsfeld or a Paul Wolfowitz?"
I'm not sure even that's a necessary question: I'm pretty sure a lot of the progressive movement would like an advent calendar of all of the people they disagree with politically.
gattsuru
1yr ago
There's also a number of nuclear history books that describe early meetings between US and Chinese military officers as very carefree and bombastic until, when wargaming, the Chinese side would give claims for how many casualties that they would proudly sacrifice in defense of their homeland, the US side would bring out then-classified nuclear calculators and give casualty estimates, and the difference between the first numbers and the second numbers would leave everyone at the table in very morbid moods.
I'm not sure how much I trust these claims -- China pledged to no-first-use in 1964, even if Americans (not unreasonably) believed the policy to have some flexibility, those early meetings necessary come from a tiny number of original sources who were more than a little biased.
Some people are already making this argument closer to its joints, but I'll separately argue that "failure" was not the central claim from the linked thread. Even if the Sexual Revolution as a whole was an improvement for a person's life, both Botond173's and WhiningCoil's frameworks apply: quite a lot of individual people bought into a promise of open sexuality with little or no negative consequence, and the modern infrastructure of the movement has not achieved and often repudiated that goal.
WhiningCoil spends more time on the ramifications -- unwed motherhood, sexual possessiveness/unwillingness to be cuckqueened, public shame over past sex work -- but the central claim is not just those have increased (though I expect he can make that argument, and while it's hard to separate the sexual revolution from other causes, it's hard to define the sexual revolution in a way that doesn't make cam work more common), but that their results were things that the sexual revolution claimed even as they occurred would be resolved, and that didn't happen. Being a single mother still sucks even if you don't think the sexual revolution impacted marriage rates, Helluva Boss uses cuck as an insult, people who claim sex work is work will still bleach their underpants (even at far smaller scales and far less readily identifiable ways!), and there's still an absolute mess of stigma and hurt feelings and drama around casual sex even if you think there's no more of it than in the 1950s.
There's a view (and not an uncommon one!) that this just means the sexual revolution hasn't won, yet; if these things are normalized or destigmatized, that at least some of the issues will fade or be solved. But there's another view, that at least some of these issues are issues on their own, and that they cause unavoidable problems. And I don't think the division is Manichean.
"Huc is considered a bit of a sensationalist" - Do we have any specific evidence of him making shit up, or exaggerating anything?
I don't think Huc clearly made stuff up, so much as he tended to relay a lot of word-of-mouth, and it's not always obvious what and how often he saw a particular matter. The "burning oil gambling" stories as an example, it's very likely the story was honestly relayed from the original Arabic work, and it's not technically a lie... but you have to use a critical eye to notice that Huc never says if he'd seen such an oil-burnt or defingered gambler.
Nor does the context always present matters accurately: there's a lot of references to thing he saw personally in "the North", and they take a rather different light when you remember that his time in "the North" was a multiyear period a well past the Great Wall, proselytizing to religious exiles after Christianity had been heavily regulated closer to central China, which is not perhaps the most representative group.
- Is there any evidence of him receiving funding and/or support from the Japanese before he wrote his book?
My understanding is that Townsend had only been shown to receive funding from Japanese groups after the publication of Ways That Are Dark. I don't think the Japanese support was why Townsend disliked China -- no small number of people did, including many Americans who also hated Japan -- and to the extent it probably did impact his later love of Japan (the change in tone to his later pamphlets is... not minor), it's not in itself unreasonable. A lot of Imperial Japan's worst policies were not exactly visible to casual inspection outside of wartime, after all.
gattsuru
1yr ago
Yeah, that's fair. Most of the stuff I've seen with LLMs either pushes for tiny models at low quantization for speed, or goes all the way to 65b on CPU for intelligence, but I'm sure there's a lot of use cases in the middle.
Why did you choose Blitzø as the trans metaphor when there’s an explicit trans character at the ranch?
I think Blitzø's name is more relevant as a deadnaming metaphor; Sallie Mae is just explicitly trans. Blitzø's not really got that much overlap with trans stuff otherwise: the extent he burned his connections to the past is more literal and accidental than figurative and intentional, and the extent he wants to reinvent himself is just to be cool rather than to actually change from who he used to be.
And that's actually a good tactic! The idea that a political goal you support has ramifications outside of your particular ideological faction is important, and using extrapolations in media can be a) a little less preachy, and b) provide reasoning to people who can't imagine or empathize with your central use case. In this case, with something that can actually happen to (non-trans) people in the real world! There's no master-class in literature happening here, by any means, but it's something a lot of activist media struggles with, and Helluva Boss handles it with something approaching finesse, such that I'm not entirely sure they did it on purpose.
Makes the “gag” about females “not” being funny in the latest episode fall flat to me.
Yeah, having it show up at the same time as the writer's name get her tagline might have been a decent shot, but it's a little too self-aggrandizing without the underlying support, and there's not really any other gag for that entire flashback sequence, beyond how awkward the villain-of-the-week can be. The extent the character is just a Westie stereotype without any particular investigation of what that actually means (like redneck in the US, it's not just a political affiliation!) is one of the areas I'd say falls more on the cringe side.
Beyond that, VivziePop's works in general kinda struggle with the general clown-nose on, clown-nose off problem: trying to be something other than a comedy (musical? drama? newgrounds flash gone wild?), and the motioning at a handful of jokes an episode. There's occasionally funny bits -- Fizz being hilariously incompetent at self-defense in s2e6 got a few giggles from me, and almost all of the dildo jokes in s2e3 hit even if a lot of viewers won't catch them on first viewing (there were people seated in those chairs!) -- but they're kinda filler for the central drama, and even as adult comedy it's a very far cry from Oglaf. It's still better than the comedians-who-aren't-trying-to-be-funny, but even if the "comedy" is a sop to funding (see Adventure Time), I can empathize with the complaint that it's definitely not self-aware enough of that.
It looks to be a reference to Journey Through The Chinese Empire, a later work by Huc. Starting at page 259:
One day, when we were passing along the road leading to Pekin, we met a party of soldiers, with an officer at their head, escorting a number of carts, in which were literally piled up a crowd of Chinese, who were uttering horrible cries. As we stopped to allow these cart - loads of human beings to pass, we were seized with horror on perceiving that these unfortunate creatures were nailed by the hand to the planks of the cart. A satellite whom we interrogated, replied, with frightful coolness : "We've been routing out a nest of thieves in a neighboring village. We got a good many of them, and as we hadn't brought chains enough, we were obliged to contrive some way to prevent their escaping. So you see we nailed them by the hand."
"But do not you think there may be some innocent among them ?”
" Who can tell? They have not been tried yet. We are taking them to the tribunal, and by-and-by, if there are any innocent men among them, they will be separated from the thieves." The fellow seemed to think the thing quite a matter of course, and was even a little proud of the contrivance.
Perhaps, what was most hideous of all in this dreadful spectacle, was the mocking hilarity of the soldiers, who were pointing out to one another with an air of amusement the contortions and grimaces of the miserable creatures in their agony of pain. If a people can exhibit such barbarity as this in quiet and peaceable times, it may be imagined of what excesses they are capable under the excitement of revolution and civil war. In the provinces now in insurrection horrible abominations must be passing.
So the core claim, of cartloads of humans nailed by the hand to the planks of the cart, is present in an original; but the soldiers did not have enough chains rather than simply forgetting their handcuffs, and rather than the soldiers bringing in a whole lot of villagers they believed (probably incorrectly!) that they had already specifically been targeting thieves (and, probably incorrectly!). More broadly, Huc brings the example forward as part of a sequence describing the severity and inflexibility of the Chinese legal code at the time, escapable only by naming someone who has committed greater crimes.
That said, I will make the necessary caveats: Huc is considered a bit of a sensationalist, and Townsend's work comes across as Chinese Cardiology (and the work as a whole is very hard to separate by his funding from an anti-Chinese group of Japanese Nationalists). Townsend also points to a Huc claim about Chinese gambling growing the vice to such an extent that some places had an axe and hot-oil bottle available, to cut off and cauterize wagered digits, and this is mentioned in the same Journey... as a story related from the 9th-century Arabic Silsilat al-Tawarkh ("Chain of Chronicles"; I've not been able to find an English version to find how well it tracks from there).
The Sexual Revolution Goes To Hell
There was a conversation a month back about the Sexual Revolution and its (Lady) Discontents, probably highlighted by this later-QC'd @WhiningCoil post:
Most people totally immersed in the mores of the sexual revolution will never be able to entertain the notion that those mores harmed them. They may look around them, at their peers, and see the damage. But their own decisions will always be above reproach, because SLAY QUEEN!!
[cw: some links NSFW, albeit more in the sense of Comedy Central late-night comedy sense. Also some media spoilers.]
Apropos of nothing, has anyone here watched Helluva Boss? 'Adult' comedy, freely available on YouTube. It stars the Immediate Murder Professionals, a trio of imps who've gained access to the living world and have offered their services to get revenge 'resolve problems' there for damned sinners who can pay. Ostensibly, the show is about the trio's new business as marginally-competent assassins, with the moral and neurotic Moxie, joyful berserker Millie, and wacky boss Blitzø ("the o is silent") going into the world and slaughtering someone. In practice, this ends up more a framing device; many episodes don't involve paid murder, and those that do it's not the actual challenge.
With a few exceptions most individual episodes instead focus more on relationships between the denizens of hell. The three main cast have that awkward mix of professional and casual common to small business (not helped by Blitz's clear desire to make a 'new family') sometimes jumping wholesale into stalking, Millie and Moxie have to juggle a marriage that's a lot more tender and reciprocal than either their parents nor Hell in general tolerates, so on.
That expands with the secondary cast. Blitz's access to the living world depends on a magical grimoire given in exchange for a transactual relationship with the demon prince Stolas, and for the first season neither are quite sure exactly how much emphasis goes on the 'relationship' in 'transactional relationship'. He also runs into a series of current or past lovers sexual partners with their complaints about him. Blitz's adopted adult daughter Loona is desperately looking for someplace to belong after a unpleasant childhood in Hell's pounds orphanages but is unwilling to risk vulnerability. Moxie has... issues with his own Family and knows that he doesn't measure up by the standards of Millie's parents. Stolas' biological daughter Octavia is desperately looking for someplace to belong while her parents go through an unusually messy divorce. Eventually a number of the Seven Deadly Sins get involved, so on.
There's a song spelling it out, diegetically as a drug trip..
As necessary disclaimer: it's gay. Really gay, even by furry-adjacent standards: there's one male/female active relationship among the main cast, and it's constantly going back to the same pegging joke. If you're a fan of the ladies, you're going to be stuck looking at fandom works or the not-on-YouTube sister show Hazbin Hotel, which does have a lesbian couple in focus. I don't know that I could call it good; while there's some decent comedic moments and fluid action scenes, there's sometimes too much emphasis on the cringe in cringe comedy, the musical numbers are hit-or-miss even if you can swing to their sometimes bizarre genre selections, and the characterization could stand to be more consistent. It's never quite Ren And Stimpy gross-out comedy, though some of the gorier fight scenes can get close, but neither is it exactly high-brow. The series as a whole has been trying to make a lot of commentary on economic and social class without serious introspection on its own assumptions, or even how that commentary it does present comes across.
((And I'm sure someone like @HlynkaCG can probably break down better about a Red Tribe take on the spirital ramifications of modern culture framing and worshipping literal demons as parallels for and paragons of modern society. Or of 'heaven' being paperwork and Minnesota Nice.))
Buuuuuuuuut because it's 'adult' and focused on relationships, a lot of it's about sex, and that part is very much written toward the id and superego of those "totally immersed in the mores of the sexual revolution". The show leads are the bisexual Vivienne Medrano and the gay Brandon Rogers, and the advertising and focus is very much down bad for exactly what you'd expect from that. That's not limited to sex -- one of the better musical numbers revolves around a two-minute long sequence of flipping the bird off to an abusive boss, culminating in a series of giant neon signs, including literal sign language for 'fuck off', the pilot has a particularly unsubtle joke about American healthcare provisioning -- but it's very much spread throughout the ethos. Of the main cast and the secondary cast, only one person (Millie) doesn't have Daddy Issues.
Helluva Boss is 'woke' in the sort of way that its authors would consider 'woke' to be a compliment. To its credit, that's at least sometimes subtle: we do some awkwardly-placed Deaf Culture-rep or a character awkwardly pointing out to his father that bisexual and gay are different things, but there's also a few trans characters (and Blitz-the-o-is-silent is probably meant as a deadnaming metaphor) or more subtle discussions about triggering trauma that you'd have to pay attention to catch. (It helps that the writers are willing to throw some on-the-nose jokes the other direction).
((It's worth spelling out that, where Hazbin Hotel discusses consent and undesired sexual violence with the characters Angel Dust and Valentino, it doesn't really feature among the reoccurring cast for Helluva Boss: the closest matters have been comedic and near-instantly resulted in violent response. Instead, the show portrays sexuality as a tool for the characters, either figuratively with many separate characters squicking out the villainous Striker to discomfort him, or in the more literal sense of skewering attackers through the skull with a motorized and pixelated dildo.))
There's actually a lot of discussion here about how modern (and thus post-sexual-revolution) norms are, in the story's setting, literally damning. "He's had four tongues inside him at once, which, like, good for him!... but he's giving off not-ok vibes" is the most clearly overt situation where the show can't quite disavow people who want to fill every hole they've got, but it can recognize that sex won't fill and often detracts from figurative ones. There's clear contradiction between more 'presentable' sexuality and less such (cw: lots of pixelated dildos, loud, se2 spoilers). STDs exist, in-universe. One of the last straws for that Two Minutes Notice song is the promotion of an entertainer as a sex object that is at best degrading and at worst invites or encourages aggressive stalkers, a topic of prolonged discourse in fandom spaces that's somewhat complicated by the number of people who literally get off from fascimiles of their body or their characters being 'used'. An early-season joke about fandom response to Loona is slightly awkward in contrast to around 15k not-always-on-model images over at e621 that I won't be linking. Though at least the character's explicitly in her twenties.
((The showrunners are probably not considering these conflicts solely as a theoretical exercise. The original voice actor for Stolas was dropped between the pilot and the first season, at the same time certain 'allegations' were going around of Totally Consensual But Also Bad things.))
A lot of the show's answer is to highlight and exaggerate the faults in 'traditional' sexual norms. Whatever sympathy the fandom came up with for Stolas' wife before her reveal -- after all, he was cheating on her! -- faltered when Stella actually appeared, less
But Helluva Boss is struggling to create and draw together a healthy sexuality after the sexual revolution, and as a response to the sexual revolution rather than just those 'traditional' norms. A good number of those criticisms are very likely inspired by personal experiences, and many viewers see and relate to the show in that framework. A bit of that is drawing very heavily from Women's Fanfic Circles of Idealized Relationships, where everybody 'really' just needs sufficient support followed by Just Admitting Their Feelings And Letting People In (something something Found Family), or is disposable and untouchably evil (and there are a lot of disposable assholes). Other parts are more serious. If relationships are increasingly likely to touch between work and play, what extent can a transactional relationship or one with disparity of power be healthy, or can such a thing ever leave those fetters behind?
That's not to say the show has answers. It's not even clear that it's entirely grappled the scope of the questions: like a lot of shows with complicated romantic relationships, there's a fan-favorite solution that's almost impressive for how much it's joked about compared to how little it's presented in any serious sense, even if only to point out where and why it wouldn't work. Some few of the protagonist's flaws are their own, but there's little space or consideration for what would be necessary to grow beyond them, or to produce a next generation that could easily exceed them.
I think it's still relevant to say that they've noticed the skulls.
gattsuru
1yr ago
EVGA's transferable warranties are nuts; it's a pity they've gotten out of the market. I'll caution that in addition to the risk of getting a bad card, you're also just going to get an older one. Be prepared to replace fans in a year or so, and recognize that you're probably going to get an earlier or lower-gaming-performance model like a 1080 SC than a 1080 Ti FTW3.
That said, there is a performance benefit to some less eye-watering intermediate upgrades that won't show up in the simple frequency check or G3DMark score, even without raytracing, especially when comparing to the simpler model 1080 SCs. It's not a huge difference, especially at the price point, but for many newer and especially VR games it's the difference between a great experience and a moderate or unplayable one.
Agreed that AMD has a number of more cost-effective options for purely gaming performance points, and at more than just the gap between the highest ends.
There are also some ML models out there that can fit into 12GB VRAM but not 8GB, though I expect anyone with that use case is already aware of them, and some rendering use cases where the difference between a 30- or 40- series card and a 10-series card is pretty (bizarrely) large.
gattsuru
1yr ago
Yeah, Birkenhead drills are far broader than just "the captain goes down with the ship" or "women and children first". In aviation, the phrase is used to not just mean a duty to passengers, but also civilians on the ground.
But I think it's important to spell out that it's not just setting a positive example, but that it did so in a way that probably saved over a hundred lives at Birkenhead, directly.
gattsuru
1yr ago
gattsuru
1yr ago
Supplies have gotten more robust, even if prices are still a little top-heavy. If you're dead set on a specific version from a specific vendor, you might have to do some shopping around or bending over backwards, but if you're just looking for a specific chip and just want a moderately competent cooler on it, you're much less likely to have a problem. Price increases after launch have been more limited to the top end and been modest even there. And nVidia has been doing some price drops or 'relaunches' shortly after initial release for a few of the 4000-series, most awkwardly the 4080 12GB 4070TI.
A lot of stores have physical inventory, albeit sometimes with per-household sales limits. MicroCenter, if you have one near you, is the obvious go-to, but BestBuys and even some WalMarts (wtf) have started stocking mid- and high-end current generation cards, and sometimes been able to keep them in stock. Note that some stores will not have all of their inventory out in the display cases, even if those display cases seem bare; I just picked up a 3060 a few weeks ago, and the MicroCenter in question only had display slots for 4000-series cards -- normally asking staff to check 'in the back' is a waste of time, but this is one exception.
That said, unless you're doing a lot of cutting-edge ray-tracing-heavy gaming at 4K or in VR, some very high-end (and given the prices, paid) rendering work, or some very specific AI/ML use cases, or have money to burn, I'd also caution against immediately going for the cutting edge. At the simplest level, there's just been a number of 'new' cards with little or no performance benefit for steep price increase, most famously nVidia's 4060's nearly at par with the 3060s, or in AMD/ATi land the 7800XT being a rounding error with the 6800XT.
That's doubly the case if you're looking at a 4070/4080 Super, which are the only cards with rumored near releases. You're probably looking at 800+ USD at the low end at MSRP, if not closer to 900 USD, and buying by a script on launch date means you'll be buying before the review embargoes fall. I don't expect them to be complete stinkers, and current SKUs aren't the worst things we've ever seen from nVidia (have you ever heard of a GTX 800-series card? had to explain the 16-/20-series bullshit?), but even if it's a reasonable wager, you're still betting a pretty sizable stack of cash on nVidia not doing something stupid.
All that said, for scripts, the tools available vary, though (given the increased availability since late 2022's crypto crash) a number would need to be updated for new cards. In essence, they're little more than a scheduled task or cronjob pinging a web page with a given search query at a specific interval; this is Python101, or if you wanted to be miserable you could do it as a curl and a painful amount of piping. It's deciding whether you want this to just ping you, or to autobuy, that's harder... but that's hard as a strategic challenge, rather than practical one.
gattsuru
1yr ago
Pointlessly suicidal demands that a captain literally go down with their ship exist, but they're outliers and often self-enforced (and sometimes overriden by other staff).
The phrase dates back to and probably originates from the Birkenhead Drill, where the ship could not float enough lifeboats for its crew and soldiers, and that as a result the officers commanded and demonstrated willingness to attempt a long (and for most, suicidal) swim to shore rather than swamp or overturn the lifeboats, resulting in a greater number of deaths.
In the modern day, (almost) all ships have enough lifeboats for an excess of passengers and crew: in these contexts, the demand is more than staff should remain until both all passengers have been disembarked (and the ship has been certainly lost, due to salvage law), closer to your preferred framework. However, spelling it out as potentially self-sacrificing is important: whether there are sufficient lifeboats and time to embark them, there remains a serious temptation for crew to save themselves while leaving their charges helpless. Passengers may not even be physically capable of the necessary actions to evacuate, nevermind have the knowledge of how or why to do so.
This is present even for other contexts, such as aviation, albeit in a lesser form. While there are exceptions for some types of incident where near-instant unsurvivable effects are likely to occur, both pilots and especially cabin crew are trained to evacuate as many passengers as their roles and positions in the plane allow before leaving themselves, and this matters.
That said, while I'm not very familiar with the 2008 Mumbai attacks, from what I've heard these causes don't really apply to hotel staff faced with spree killing terrorists.
gattsuru
1yr ago
At the risk of Godwinning where the comparison is genuinely hyperbolic, “if only the Fuher knew”?
I’m not intimately familiar with Aussie politics, but I’m familiar enough to give a list of recent cases where stupid bullshit safetyism ended up with friends in high enough places for politicians to happily stand by them, whether that be the various recent fishing bans (ban shark fishing to prevent shark attacks!), the Adler shotgun, or the obnoxiously early ADS-B mandate.
I mean, it might be worth a shot, but the assumption no one has taken that try simply because it hasn’t changed already is hilariously naive.
gattsuru
1yr ago
... depends on what you're asking. We don't have data, and such data's probably impossible to get, but there's some stuff you can reasonably guess.
Does the extent of the current policy put off more advertisers or watchers or creators than a milder one would? Probably! Does a content policy doing some of these rules probably make them more advertisers/watch/creators than having absolutely zero would? Also probably, given that above any natural economic motions it'd attract the same sort of 'oversight' that drove Media Matters to fuck with Twitter recently.
Does any of that matter, when we already know every payment processor outside of Russia will drop them like a sack of potatoes?
gattsuru
1yr ago
Sam Bankman-Fried, of FTX fame.
gattsuru
1yr ago
gattsuru
1yr ago
There's a lot of controversy and reputation about the man in economics, but I don't really have a good level of confidence in evaluating whether he's right or wrong on those.
Summers has a lot of pieces like this or this in his history. It's more generally a side effect of his state-first thinking, and it's not even always wrong, but it's very much the normie version of "AI Safety" as about jobs programs or small-scale disruption. Maybe his personal experience at Harvard will stop him from advocating for RLMFing LLMs in loops; I'm not confident on that one.
Makes him a very high-profile and high-reputation version of the only opponents of things I like can be Luddites view.
gattsuru
1yr ago
We have reached an agreement in principle for Sam Altman to return to OpenAI as CEO with a new initial board of Bret Taylor (Chair), Larry Summers, and Adam D'Angelo.
…
Larry Summers
Fuuuuuuuuuuuck.
gattsuru
1yr ago
The theory is that a) there's so much power and money on the table here that someone doing something world-changing is unavoidable, and b) that the early design stages of such a world-changing system will have massive impact on whether. There are ways to argue against either or both of these assumptions, or to point out separate issues.
((Some of which I agree with: even accepting those propositions, these guys are demonstrably putzes when it comes to actually persuading or planning.))
But 'it's like a religion' isn't a very meaningful claim.
gattsuru
1yr ago
I'm not disinterested but I'm also not sure how much ore there is left to mine.
I think there's a lot of space unexplored; I'm just not sure what part actually matters.
There's a lot to be said about whether utilitarian philosophy demands or can't avoid paperclipping behavior, a lot to be said about whether paperclipping behavior requires utilitarian perspectives or underpinning, a decent amount to be said about what extent modern ML uses goal functions and how much these meaningfully overlap with utilitarianism (if at all), and some stuff to be said about whether Yudkowsky gives off bad vibes / the only thing LW is "interested in is recognition for being very smart."
But some of these matters are far more meaningfully debatable as matter of fact or disprovable theory than others, and not just in the sense that appeals to messages sent to an account you don't want named or linked to your current one are hard to meaningfully discuss in an honest way.
gattsuru
1yr ago
This seems... weird, as an explanation, and given my expectations for the NYT may reflect more what one party has fed to the reporter than the real facts on the ground.
The Toner paper in question is here, and there's wayback machine version dating back to Oct 24th. The closest bit I can get to the description from the NYT piece is the section where :
While the system card itself has been well received among researchers interested in understanding GPT-4’s risk profile, it appears to have been less successful as a broader signal of OpenAI’s commitment to safety. The reason for this unintended outcome is that the company took other actions that overshadowed the import of the system card: most notably, the blockbuster release of ChatGPT four months earlier. Intended as a relatively inconspicuous “research preview,” the original ChatGPT was built using a less advanced LLM called GPT-3.5, which was already in widespread use by other OpenAI customers. GPT-3.5’s prior circulation is presumably why OpenAI did not feel the need to perform or publish such detailed safety testing in this instance. Nonetheless, one major effect of ChatGPT’s release was to spark a sense of urgency inside major tech companies.149 To avoid falling behind OpenAI amid the wave of customer enthusiasm about chatbots, competitors sought to accelerate or circumvent internal safety and ethics review processes, with Google creating a fast-track “green lane” to allow products to be released more quickly.
This result seems strikingly similar to the race-to-the-bottom dynamics that OpenAI and others have stated that they wish to avoid. OpenAI has also drawn criticism for many other safety and ethics issues related to the launches of ChatGPT and GPT-4, including regarding copyright issues, labor conditions for data annotators, and the susceptibility of their products to “jailbreaks” that allow users to bypass safety controls. This muddled overall picture provides an example of how the messages sent by deliberate signals can be overshadowed by actions that were not designed to reveal intent.
A different approach to signaling in the private sector comes from Anthropic, one of OpenAI’s primary competitors. Anthropic’s desire to be perceived as a company that values safety shines through across its communications, beginning from its tagline: “an AI safety and research company.” A careful look at the company’s decision-making reveals that this commitment goes beyond words. A March 2023 strategy document published on Anthropic’s website revealed that the release of Anthropic’s chatbot Claude, a competitor to ChatGPT, had been deliberately delayed in order to avoid “advanc[ing] the rate of AI capabilities progress.” The decision to begin sharing Claude with users in early 2023 was made “now that the gap between it and the public state of the art is smaller,” according to the document—a clear reference to the release of ChatGPT several weeks before Claude entered beta testing. In other words, Anthropic had deliberately decided not to productize its technology in order to avoid stoking the flames of AI hype. Once a similar product (ChatGPT) was released by another company, this reason not to release Claude was obviated, so Anthropic began offering beta access to test users before officially releasing Claude as a product in March.
Anthropic’s decision represents an alternate strategy for reducing “race-to-the-bottom” dynamics on AI safety. Where the GPT-4 system card acted as a costly signal of OpenAI’s emphasis on building safe systems, Anthropic’s decision to keep their product off the market was instead a costly signal of restraint. By delaying the release of Claude until another company put out a similarly capable product, Anthropic was showing its willingness to avoid exactly the kind of frantic corner-cutting that the release of ChatGPT appeared to spur. Anthropic achieved this goal by leveraging installment costs, or fixed costs that cannot be offset over time. In the framework of this study, Anthropic enhanced the credibility of its commitments to AI safety by holding its model back from early release and absorbing potential future revenue losses. The motivation in this case was not to recoup those losses by gaining a wider market share, but rather to promote industry norms and contribute to shared expectations around responsible AI development and deployment. Yet where OpenAI’s attempt at signaling may have been drowned out by other, even more conspicuous actions taken by the company, Anthropic’s signal may have simply failed to cut through the noise. By burying the explanation of Claude’s delayed release in the middle of a long, detailed document posted to the company’s website, Anthropic appears to have ensured that this signal of its intentions around AI safety has gone largely unnoticed. Taken together, these two case studies therefore provide further evidence that signaling around AI may be even more complex than signaling in previous eras
Yes, this is weird writing, in the sense that it's (a little) odd for someone to praise their market competitor so heavily, and it's also a trivial thing to get that bent out of shape about either way, but we're talking about a bunch of self-considered weird auteurs; it'd be less believable to not have some tyranny of trivial disagreements involved.
Ultimately, the option is a Morton's Fork between the dead and the forsworn. "Why" they bent the knee at that level isn't particularly interesting: Clement as the office and officer of one of the most prestigious BigLaw firms would be gone either way.
((Indeed, Clement did take the option of earning a
mediocrenot gold-plated living.))But that's not very interesting; the number of political groups who can't be bought at any price is pretty low, and they're not exactly the most effective or most palatable ones. Step back one point to see why these groups could bring this credible and this serious of a threat, and you notice that it's something far broader and longer-lasting driving things.
More options
Context Copy link