PendingKetchup's recent activity
-
Comment on Why do LLMs freak out over the seahorse emoji? in ~tech
-
Comment on Timeout when connecting to a local webserver through the internet, but only on WiFi in ~comp
PendingKetchup LinkI would look at the AP, since it is another box between you and the destination on the wifi path, and probably has a web UI and some ability to firewall. What does it think its own IP is? Is it...I would look at the AP, since it is another box between you and the destination on the wifi path, and probably has a web UI and some ability to firewall. What does it think its own IP is? Is it somehow doing triple NAT?
-
Comment on The Buff Scammer, isolation, and the male loneliness epidemic in ~life.men
PendingKetchup Link ParentI think I think of power gradients as kind of like the gradients of multuvariable calculus. If you have like rich white woman CEO over here, and a poor homeless noncitizen Hispanic man over there,...I think I think of power gradients as kind of like the gradients of multuvariable calculus. If you have like rich white woman CEO over here, and a poor homeless noncitizen Hispanic man over there, the first might have a higher "absolute" level of power than the second, but if you look at the derivative just along the gender dimension, that component points the other way. That local gradient might then be relevant if you are trying to think about gender-related aspects of their interactions, like who would be more likely to be taken to task for not giving enough attention to their children. I'm not really sure what current thinking is on who should be believed more if each simultaneously accuses the other of sexual assault, because you've put the "believe women" maxim up against the rest of the body of critical theory. I know you aren't meant to do a sort of "oppression Olympics" or start assigning point values to identity characteristics, but I don't really know what you are supposed to do instead.
Sexual exploitation indeed doesn't only happen down the gender gradient. Any power imbalance, for any reason, creates an opportunity for sexual exploitation. If people are going around thinking that is only a man thing, they are wrong. But that also means that you should take power dynamics other than gender into account when thinking about sexual assault.
Even if you don't feel more powerful as a man than women who are your peers, the theory says that the local gradient wherever you are is still "more woman -> less power", so if you were everything you are but also a woman, you would be predicted to have additional problems. I suppose it's possible you've found a place where that isn't actually true, but the theory also says that power disparities are much harder to see from the high side than the low side: the work of thinking about and dealing with them generally gets put on the people with less power. So you might want to check with the women peers to get their perspectives.
-
Comment on Bitnami’s August 28th bombshell: The end of free container images as we know them in ~tech
PendingKetchup LinkIn the middle of the article it lapses into "increasingly breathless Markdown headings and lists" style. We have "Dependency Hell Multiplied" and "Legacy Repository Limbo" and a bunch of sections...In the middle of the article it lapses into "increasingly breathless Markdown headings and lists" style. We have "Dependency Hell Multiplied" and "Legacy Repository Limbo" and a bunch of sections with three bullet points, where each bullet has its own boldface key point, and the negativity and fear is relentless.
Does everybody on Medium write like this, when they write these blog posts that are really about how their product solves $problem_of_week? Is that where the robots get it? Or was this section on "what are the potential problems this could cause my readers" maybe synthesized to spec to make writing the post faster? Or are actual people starting to write like this now, having seen too much of it?
It's not that I begrudge a corporate blogger a text generator, but I've seen this shape of text too much while being bullshitted by an instruct-tuned model that I have asked for something which it is really incapable of delivering. It is is maybe trying to make up for in formatting and stridency what it is failing to dredge out of its few billion weights in actual useful information, and hide the fact that it is just restating the obvious from whoever is being paid a piece rate to rank the responses. So when I see it in the wild it sets off my BS detector, and I feel inspired to try and reboot whoever is talking to me.
-
Comment on The Buff Scammer, isolation, and the male loneliness epidemic in ~life.men
PendingKetchup Link ParentI think this is just a failure to reconstruct a normative set of rules for social interactions, given current theories of consent. There's a model in which no one may physically touch anyone at...I think this is just a failure to reconstruct a normative set of rules for social interactions, given current theories of consent.
There's a model in which no one may physically touch anyone at any time without verbally asking for and receiving permission from someone who is under no particular pressure to say yes. This might be correct! Maybe extending a hand to shake in a business context should be forbidden, and we actually need to start with something like "May I shake your hand? You are completely fine to say no, I won't be upset."
In practice in most spaces we are still using a lot of implicit negotiation, or non-negotiation, and people are still sort of poking each other and seeing if anyone complains, tapping each other's shoulders, leaning in to do that cheek kissing thing they do in Europe, and so on. It might be possible to make that workable under a developed theory of consent, if it essentially rises to a language that everyone involved knows, but it also might just all be wrong and need to change.
It's easier to see the problem when it is flowing down the power gradient instead of up, and e.g. the person who makes twice as much money at work and does not have the state taking a malicious interest in their guts is the one who did not in fact have permission to punch the other one in the arm for a bad pun. And if we persist in having any acceptable social interactions that are meant to be initiated and then accepted or refused, then maybe some things really are only wrong when going down the power gradient, because the person at the bottom is under pressure to permit them that the person at the top is not.
-
Comment on The Buff Scammer, isolation, and the male loneliness epidemic in ~life.men
PendingKetchup Link ParentIt's going to be even lower because they won't all be unrelated. A small fraction of one group is terrorizing a large portion of everyone else.It's going to be even lower because they won't all be unrelated. A small fraction of one group is terrorizing a large portion of everyone else.
-
Comment on The Buff Scammer, isolation, and the male loneliness epidemic in ~life.men
PendingKetchup Link ParentOne missing piece here is the power landscape. While stereotyping and criminalizing groups of people is never OK, it hits different when it is "punching down", so to speak. You can point to a...But I could see how at a much larger scale how that could be hurting boys and young men. We're stereotyping and even criminalizing them from the start. It's not OK when we do that to black youths. It's not OK when we do that to Muslims. I don't think it's OK here.
...Right? Hmm.
One missing piece here is the power landscape. While stereotyping and criminalizing groups of people is never OK, it hits different when it is "punching down", so to speak. You can point to a statistic as potentially explanatory for a cultural attitude without laying down another brick in the wall of the oppression of men, because in the society we currently inhabit, for all that they do have their own specific problems, "men" are not marginalized.
But also, in this system of thought, I don't think collective guilt is meant to be real, because it's not a rules-based, deontological system that works on concepts like "guilt". It's a consequentialist, "harm"-based system. So while a member of a dominant group is obligated to be aware of, say, how walking near a member of an oppressed group might make the other person feel, it is not because of anything wrong with or collectively bad about the members of the dominant group, and it is entirely about the moral imperative to protect the members of the marginalized group from further negative impacts.
Black people in the US are held to be owed reparations, as I understand it, not because of the collective guilt of their enslavers, but because of what was taken from them.
Since it's all based on outcomes, there is no correct set of rules to follow. It's not even, really, a matter of being obligated to think about how you might affect someone negatively and take "reasonable" steps to avoid it. It's a matter of being obligated to succeed in not negatively impacting them, by either walking too close or too far away. You thus obviously can't win all the time, and nothing you personally do can ever be enough. This accurately reflects how no individual member of a dominant group is going to personally dismantle the enclosing system of oppression.
-
Comment on The Buff Scammer, isolation, and the male loneliness epidemic in ~life.men
PendingKetchup Link ParentI think there might be a stereotype-generating effect of large groups going on here. If you take dozens of different people, and get one fragment of a desired gender role from each, and AND them...When I think about implicit expectations versus explicit asks, I think the most poignant modern phenomenon is the concept of "the ick". In theory, the definition is some random, inconsequential behavior that inexplicably causes a woman to lose attraction to an otherwise attractive man.
In practice, it's just overwhelmingly an enforcement of strict gender roles. A man gets too excited about something he's passionate about? Ick. A man cries because he got emotional about something personal to him? Ick. A man skips, or eats an ice cream cone, or carries a hand bag or takes baths? Ick.
They're not all enforcing traditional masculine gender roles, but it's hard to read any compiled list of icks as a whole and read them as a prescription for a very specific type of traditional masculinity, one which explicitly does not include sharing any emotions except for anger and lust.
I think there might be a stereotype-generating effect of large groups going on here. If you take dozens of different people, and get one fragment of a desired gender role from each, and AND them together, you're going to get an extremely tiny and plausibly empty region of gender space. If each had kept just one piece of a received "man" gender as something they personally actually like, you can end up reconstructing the whole thing even when no one individually wants that.
Nobody can be, or should try to be, desirable to everyone.
I think you get this as a cultural phenomenon because you are throwing strangers together and expecting them to date. If you have a good read on someone to start with, you're not likely to be surprised and prompted to re-evaluate them by their consumption of ice cream or by them having a good cry. And if whoever it is is someone actually in your community, who won't solipsistically cease to exist if you stop dating them, then deciding you don't actually want to date them anymore has much less of the character of discarding something disgusting than the word "ick" would imply.
-
Comment on US FBI readies new war on trans people in ~lgbt
PendingKetchup (edited )LinkDoes everybody have their tee shirts ready for the war? If the FBI wants to gin up a terror network based on the concept of being transgender, they will have to do a lot of ginning up. There are...Does everybody have their tee shirts ready for the war?
If the FBI wants to gin up a terror network based on the concept of being transgender, they will have to do a lot of ginning up. There are only so many Zizes. There are a lot of people with a lot of experience holding space for their friends, which does not a conspiracy to commit politically-motivated violence make.
-
Comment on Musings on "Developer Mode" in ~comp
PendingKetchup Link ParentI think we could benefit from some overthinking here. On one level, it seems perfectly natural not to bother the normal users with all these widgets and gizmos that they won't understand, can get...I think we could benefit from some overthinking here.
On one level, it seems perfectly natural not to bother the normal users with all these widgets and gizmos that they won't understand, can get them into trouble, and for which they probably can't formulate a thought that would constitute a problem that one of them could solve, lacking the right mental model of the system. Since the tools still need to exist for developers to develop, it's obvious there should be a setting to tell the program if you are a normal user or a developer.
But on the other hand, this is a decision someone made for it to be like this. Rather than designing a UI that can progressively expose the full power of the underlying system by progressively educating each user and leading them along a journey to mastery, which is hard, someone decided to build an "I am a member of a category of people who are responsible for their own education and whom you may legitimately confuse" switch instead.
That might be better UI when you consider a fixed group of people coming in with a fixed set of problems as new users, some of whom are developing stuff and most of whom are not, because it's nice and simple and still can handle advanced operations for people who already know they need them. But this UI is not effective at educating its users to bring them across that gap it creates; it relies on something else (a university degree, a "For Dummies" book, the plugin-authoring tutorial in the written documentation) to do that instead.
I don't think one approach is better than the other from internal principles of UI design: it is a political question of what you want the software to do to and for its users.
-
Comment on Musings on "Developer Mode" in ~comp
PendingKetchup Link ParentI'm also a person who uses browser dev tools professionally (although not as my main project). I know how they are useful for debugging while doing web development: you can fiddle with your...I'm also a person who uses browser dev tools professionally (although not as my main project). I know how they are useful for debugging while doing web development: you can fiddle with your styling until it looks exactly right, you can see what DOM you really have and what happened to your network request you were expecting, and so on.
What I'm aiming at is the seemingly commonly accepted notion that there is a kind of person called a "developer", who these tools are for, and that if you have to ask, as it were, that's not you. I produced examples of cases where browser dev tools would be useful for solving non-development problems in order to illustrate the pattern in the set of use cases that the browser makers are actually targeting.
Maybe this just reflects the shape of the world right now: the browser makers probably have telemetry showing that almost all dev tool use is during business hours, by people who use them a lot, on specifically localhost and one or two sites that presumably belong to their employer. They have measured and characterized a target audience that really does have its own coherent set of needs and arguably is typologically distinct from everyone else in the dataset, and they call them "developers".
But making and reinforcing this distinction is an important part of the politics of the artifacts here, and it is that politics I want to discuss. I didn't learn how to write a web page in a web developer training program, I learned how to write a web page from a little hardback children's book, read by flashlight in a linen closet, when the political possibility of an Internet built by just anybody was still alive. If you give that book to a kid today, do they meaningfully join the ranks of "developers", now that they have CSS-related problems? I think the question is silly and the categorical, identity-based thinking behind it, and behind the artifacts that suggest it, is unnecessarily restrictive.
I totally misread/didn't read the WikiHow article, but even ChatGPT seems to know what a "developer mode" would be like if it had one. I think that the fact that this is understood as a role people have (being a developer), and not a thing people do (developing something), is a political decision, and should be recognized as such.
-
Comment on Musings on "Developer Mode" in ~comp
PendingKetchup Link ParentI thought that example output looked suspicious. I remember that news going around now, and this is obvious if I actually read the guide I linked for more than 10 seconds. I think that's even more...I thought that example output looked suspicious. I remember that news going around now, and this is obvious if I actually read the guide I linked for more than 10 seconds.
I think that's even more interesting, though. It doesn't have a developer mode, but the model can produce the flavor of one on demand.
-
Musings on "Developer Mode"
Peruse this relevant meme. It depicts the magical transformation that occurs at the moment one taps the Android build number for the seventh and final time, as the arcane ritual transforms one...
Peruse this relevant meme. It depicts the magical transformation that occurs at the moment one taps the Android build number for the seventh and final time, as the arcane ritual transforms one from a chill dude in a business suit into that powerful, shadowy figure known only as "a developer".
It's a joke, obviously, but only half a joke. The "You are now a developer!" message that the developers at Google programmed your phone to display, when it grants you this set of powers that Google permitted them to program it to grant you, is doing something in the model of the world that its authors live in.
"Developer mode" isn't just for Android. The browser you are reading this in has a little panel you can open to inspect or adjust the content of the page. It's useful for things like composing humorous screenshots, deleting annoying ads, and downloading images and videos, but it's called the "Developer Tools", a set of tools defined not by what they do but by who they are intended or imagined to be doing it for. Discord has not only a "developer mode" that lets you get the permanent identifiers for messages, but also additional developer-exclusive functions that are activated by enabling the Electron developer tools and injecting code to set the
isDeveloperflag. Windows has a Developer Mode. ChatGPT ~got one for some reason~ has a popular jailbreak based around convincing it that it has one. This notion that a special class of people called "developers" exist, and that they must or should be afforded extra power in our society's digital spaces, is woven into the structure of the digital environment.Why is it like this? Big Tech doesn't give any power for free. Is it something their labor force of developers demands to be able to grant to their counterparts outside the company? Is it a Ballmer-Doctorow gambit of courting programmers as potential business customers by temporarily empowering them, before they start putting up the prices on the code signing certificates? Is it to distract and mollify hackers, to keep them from seizing similar powers in a more destabilizing way?
Is there any truth to the notion that "developers", independent of whether or not they are currently testing or programming something, are a class with different needs and rights from normal humans?
17 votes -
Comment on HELP: Suddenly seeing a huge influx of ethernet devices on my network in ~tech
PendingKetchup LinkIt looks like you have a lot of entries with the same MAC address and different IPs. And over on the right the allocation type is "static", which is weird if the theory is that something or a...It looks like you have a lot of entries with the same MAC address and different IPs. And over on the right the allocation type is "static", which is weird if the theory is that something or a bunch of things are showing up and getting a bunch of IPs by DHCP, because that would be called "dynamic".
Are you sure you didn't somehow configure the router to assign dozens of IPs to this MAC? Or is the "static" column lying to you for some reason?
Do you maybe have a device or a VM that keeps releasing and renewing (or not releasing and then still trying to renew) its IP? If the MACs are all the same, and the "static" column isn't to be believed, this could be one device asking for IPs repeatedly, and getting a different one each time (either because the DHCP server is not very clever, or because the device is using the Client ID option to get a bunch of different leases on purpose).
If you fire up Wireshark and tell it to filter for DHCP traffic, you might get some insight into what is being asked by the device. You could also see how frequently it is asking for new addresses and how practical it would be to wander around unplugging things until it stops.
-
Comment on <deleted topic> in ~arts
PendingKetchup Link ParentIt sounds to me like this comes from a couple sections. One describes the material conditions leading to the production of "slurry": And one describes Rosalind Krauss's "post-medium condition":...It sounds to me like this comes from a couple sections. One describes the material conditions leading to the production of "slurry":
An oversaturated attention economy means you also have a lot of pressure for one piece of “content” to do double, or triple, or sextuple duty across platforms, to justify the original investment of labor, with a corresponding squeeze on the thought that goes into each specific form that something takes.
And one describes Rosalind Krauss's "post-medium condition":
the idea that artists had stopped thinking about the specificity of what worked best in one medium and not another
The decision to present a transcript of a video on equal footing with the video itself, the idea that people could come by and read the transcript as a morally equivalent experience to watching the video version, the proposition that everything really important about the work can be, or ought to be able to be, presented independently of the nature of its medium, is, I think, what the article is attacking.
When you try to adapt a story from one medium to another, like a book to a movie, you can end up with a cinematic masterpiece, but only if you give the adaptation sufficient artistic attention on the terms of its new medium, and are willing to sacrifice the aspects of the original that don't translate well and add new aspects to fill spaces the new medium creates. When you're done, if it's any good, you really have made a new, transformative work, no matter what Disney's copyright lawyers might tell you.
When you take a video, and then present its transcript as if it were worth reading as an article, and you also take the audio and publish it in a podcast feed, and you clip out the best bits and crop them to portrait and put them on YouTube Shorts, you end up making a lot of mediocre adaptations that haven't had the attention put into them to make them good. And when the pressures of this process feed back onto the original work---making sure all the action of the original video is in the middle of the frame so it shows up in the short, and that the dialogue is all in paragraphs that scan well in the transcript, and so on---you end up making a mediocre original in the service of producing these adaptations.
I think you're right that this is in tension with "accessibility". In one sense, the relentless drive for making one's work able to be accessed by the largest possible audience, on all possible platforms, whatever they happen to also be doing at the time, while one is merely a single person and not actually skilled in all those media, is a key motivating factor driving the production of low-quality adaptations and excessively adaptable originals. This notion of accessibility is in legitimate tension with artistic intent. It's a defensible artistic statement to produce work that is deliberately hard for people to access: you have to come to the desert to see the sculpture in the cave, no photographs of the light show will really capture what's interesting about it, and so on. It's also the sort of statement only an artist in a position of power can make. A threatened or commodified artist needs to take your preference for reading over watching seriously, and cater to it as much as they can, even if that divides their limited resources.
But in another sense, an artist is arguably wrong to use their power to refuse accessibility, when we're talking about disability accessibility, even when that would feed back into and change the original. The W3C recommendations on alt text prescribe that designers should "avoid text in images", in service of making their web sites accessible to people who need to use screen readers to interact with the web. It's not clear that "but I wanna have text in images, for art reasons" is a valid defense here. It certainly won't argue you into an AAA WCAG rating. And it's also not clear that anyone is really rightly empowered to be making a distinction between, say, people who need a transcript and people who merely want one. Or that the transcript being, by Word of Authorial God, a lesser or wrong way to experience the work, isn't the sin of ableism.
-
Comment on I had an idea for a Crusader Kings, but about rich families in Victoria-Modern Era. What could go wrong? in ~games
PendingKetchup LinkYour technical problems are solvable. You might not want 10k characters kicking around as Godot nodes, but even though Godot isn't an entity-component-system engine you could bolt one on and write...Your technical problems are solvable. You might not want 10k characters kicking around as Godot nodes, but even though Godot isn't an entity-component-system engine you could bolt one on and write a "character system". Or you could skip the framework and write your own little module for doing "character physics" quickly with all the character data compactly stored in the Big Character Array. Then you could use Godot nodes as UI and instantiate stiff to show just the characters the user is looking at. Then you dump the giant array to disk and that's basically the save game. You can fit 10k characters of 1k bytes in 10M of memory or disk, which is easily manageable.
I think your bigger problems are going to be with theme, tone, and what you are trying to say with your game.
There are two kinds of Crusader Kings players. The players you want are your Jack de Quidts, who play these games with (or to develop) an understanding that feudalism was bad. There's no better indictment of a system than experiencing firsthand how its structure is making you, a fundamentally good person, really really really want to torture innocent people in your dungeon, so you can learn their secrets, so you can blackmail your neighbor into swearing fealty to you, so you can shake off the yoke of your own overlord.
The players you don't want, they log thousands of hours and still somehow think that if there was a Count of San Francisco, it would be them, and it would be pleasant. You do not want these dudes in your community. They are bad news all around.
Your game needs to be designed to communicate that, while the mechanics of "owning global company empires" present interesting problems, it's a fundamentally bad idea in real life. Put the player's moral principles in strong and obvious tension with the demands of optimal play. Put another character on screen a few times so the player has a chance to grow to like them, and then pull in that character, not a random one, to fill the slot for the event where the player has a great opportunity for trechary. Make sure that, by the time the player is on top of the world, there's not a lot of world left, and the rising seas are coming for their private island. Every time they see the "You Win" screen, you want the player asking themselves "But at what cost?". Everyone going in is going to want to be the kind-hearted local chain owner: your job as a game designer is to set up systems that force them to compromise those principles, thus communicating why those guys are so thin on the ground.
-
Comment on Google has released data on how much energy an AI prompt uses in ~enviro
PendingKetchup Link ParentBut video takes more, I think. Maybe text is worse with the truly huge models, but you can push acceptable text out of a 20b model, and acceptable video out of a 14b model, but the text model can...But video takes more, I think. Maybe text is worse with the truly huge models, but you can push acceptable text out of a 20b model, and acceptable video out of a 14b model, but the text model can run at about reading speed on equipment you can buy, while the video model runs noninteractively at like 2 minutes per 5 seconds of output. So call that 2 minutes microwaving per plausibly useful video.
-
Comment on Google has released data on how much energy an AI prompt uses in ~enviro
PendingKetchup Link ParentI'm not sure the chipset on a server is an appreciable power draw? The estimate here is probably good to a factor of 2, which is probably good enough for deciding whether to send another marginal...I'm not sure the chipset on a server is an appreciable power draw? The estimate here is probably good to a factor of 2, which is probably good enough for deciding whether to send another marginal query to the model or not.
If you accept the framing that you are likely to have a median query (with a median chance of cache hits?) and that what really matters is marginal query cost (and not the effect on their internal metrics that could lead them to do more training runs or design more AI products).
-
Comment on Germany legal case alleging adblockers violate copyright in ~tech
PendingKetchup Link ParentHow does your unidirectional theory of political change explain things like Margaret Sanger's victory over the Comstock Act?How does your unidirectional theory of political change explain things like Margaret Sanger's victory over the Comstock Act?
-
Comment on What heritability actually means in ~science
PendingKetchup LinkThe article seems pretty good at math and thinking through unusual implications, but my armchair Substack eugenics alarm that I keep in the back of my brain is beeping. Saying that variance was...The article seems pretty good at math and thinking through unusual implications, but my armchair Substack eugenics alarm that I keep in the back of my brain is beeping.
Saying that variance was "invented for the purpose of defining heritability" is technically correct, but that might not be the best kind of correct in this case, because it was invented by the founder of the University of Cambridge Eugenics Society who had decided, presumably to support that project, that he wanted to define something called "heritability".
His particular formula for heritability is presented in the article as if it has odd traits but is obviously basically a sound thing to want to calculate, despite the purpose it was designed for.
The vigorous "educational attainment is 40% heritable, well OK maybe not but it's a lot heritable, stop quibbling" hand waving sounds like a person who wants to show but can't support a large figure. And that framing of education, as something "attained" by people, rather than something afforded to or invested in them, is almost completely backwards at least through college.
The various examples about evil despots and unstoppable crabs highlight how heritability can look large or small independent of more straightforward biologically-mechanistic effects of DNA. But they still give the impression that those are the unusual or exceptional cases.
In reality, there are in fact a lot of evil crabs, doing things like systematically carting away resources from Black children's* schools, and then throwing them in jail. We should expect evil-crab-based explanations of differences between people to be the predominant ones.
*Not to say that being Black "is genetic". Things from accent to how you style your hair to how you dress to what country you happen to be standing in all contribute to racial judgements used for racism. But "heritability" may not be the right tool to disentangle those effects.
I thought the explanation was pretty good, when it was really there an not just tell G me to ask a model and hope it explained itself correctly. L
What these things really are more than anything else is ways to slide around in word vibe space, so it makes sense that one would output emoji by vibrating with the energy of emoji-ness and of whatever thing was supposed to be being put in emoji form, and then trying to express that.