wraelk
30m ago
Wedding planners are essentially connecting couples with event vendors in a way that's aesthetically pleasing, and while the couple goes away the vendors stay in their "toolkit" to be used for the next couple. The planner is going to be generally capable of creating another copy of whatever events they've done in the past, so if a couple likes a wedding the planner previously did, they will be interested in hiring the planner. It's also pretty easy to evaluate: did you or did you not think the wedding looked nice?
A matchmaker is connecting two people with each other, and both sides of that go away: if you see them pair up Matthew and Laura, that doesn't mean that they can produce another Matthew on demand, or that if they could Matthew would be a good match for the next customer. And because it's symmetrical and relational, it's kind of weird to evaluate: the ideal matchmaker for me would produce someone I think is way better than I would have been able to otherwise get, but does that mean he's a bad matchmaker for that person, or are we just relying very heavily on subjective criteria (so she also values me higher than she would have been able to otherwise get)?
Network effects apply differently here too: vendors are happy to give a planner their contact information, because worst case scenario they just don't get any jobs from the planner. But the first person signing up to a matchmaker is getting access to a completely empty Rolodex and a long wait, and unless the matchmaker gets very lucky the second person is going to have a similar problem. Before you start getting any real matching capability they're going to have to have a lot of people paying for nothing.
I guess they'd just have to work for free (while spending for advertising to build out the list) for a while, but those people are also going to be dropping off their contact list as they play the field and occasionally find something on their own while waiting for a call.
wraelk
23hr ago
Sure, but would we expect that not to have been the case before?
"The military still has non-zero status" is totally compatible with "the military's status has significantly declined"
wraelk
23hr ago
Yeah that's fair, it's certainly hyperbolic to say NOTHING supports the military, but I think it's massively down and accelerating downward.
- Michael Bay definitely has some pro-military stuff, but it's certainly headed downhill: Ambulance has a down-and-out Marine protagonist, 6 Underground has vigilantes doing what the military won't (with a PTSD sniper), 13 Hours is all private and foreign militaries, and the last Transformers has both good and bad military.
- Top Gun 100%, but it's also Hollywood failing to create anything new and rehashing a 1986 movie. I guess props to them for not subverting the themes of the original movie, good chance that would have cost them at the box office but I don't know why I'd expect that to stop them.
- Battleship is a 14 year old movie.
These are all ongoing trends and it's not like there's a cliff after WW2, but the original premise was that we have damaged all status-granting institutions for men and I definitely think the public perception of the military is damaged.
(perhaps besides the military?)
I mean, the status of serving in the military is clearly not what it used to be, at least in America. People will still go "thank you for your service" and let them board airplanes earlier, but:
- Every military action America takes is seen as evil by a pretty sizable subset of the population, so even if they're not willing to go after the individual soldier, there's always kind of a backhanded "you're a tool for the evil empire"
- Secondary to this, wars are a lot less total than they used to be, so actually fighting the war has you second-guessing what the eventual perception of your actions are going to be. WW2 and back, "blowing up this thing that belongs to the enemy" was uncontroversially good because the bad guys were the bad guys, now it's "let's hope this doesn't end up in the ICC".
- Media portrayal of the military has gone way downhill: sometimes they'll at least get to showcase the cool equipment, but "PTSD-haunted danger to himself and others" is a way more common trope than "dashing young lieutenant". The upper echelons are portrayed as out-of-touch plot barriers, and the organization as a whole is often portrayed as wildly incompetent (although to be fair a lot of this is just necessary to get the plot to work, i.e. every zombie movie). I can't remember the last time I've seen "the military comes in and saves the day" as a plot device, though admittedly my media consumption is unusual so I could have missed a few.
- Online sentiment tends to land somewhere between focusing on the idiocy of the lower ranks (Hellcats at 183% APR, Jackass stunts) or complaining about how people working for the military are enabling genocide/oppression/what have you.
Whether these things are improvements in how we think about the world is debatable (certainly there was a lot more nuance to war than perhaps the public was appreciating, and we do have rather a lot of PTSD-haunted veterans), but it's clearly a loss in status to the people serving.
wraelk
2d ago
Not a particularly good week for Parsefinder: I'm really bogged down in class and subclass handling, which has a lot of weird corner cases I need to get into. I realized there's a weird spectrum of what defines a subclass versus just a class feature with selections:
- A class like Barbarian gets one choice, which gives multiple features over the course of character progression. Fine. That's easy.
- A class like Investigator gets one choice but gets all its benefits up front. Still pretty easy.
- Thaumaturge gets to pick three implements over the course of its levels, and also gets to pick ones to upgrade (you'll get the basic effect of all 3, two will get a second effect, and one will get a third effect). Weird in that the selected subclasses may mean different things in terms of abilities from the order/level they're picked, but not that bad.
- Commander picks a half dozen tactics, which it can change at will, but they're class features. They're not subclasses but they aren't really distinct from subclasses in terms of how they're worded, just how they work mechanically
- Champion has up to four subclasses, depending on how you look at it: the Cause that determines their reaction, a choice of Blessing that's one of three small buffs, a choice of focus spell, and a holy/unholy sanctification. They're all single choices with a single effect, but all have follow-up feats that are dependent on them, and I'm not sure which ones I'd consider subclasses and which ones I wouldn't: Causes clearly are, but the others are much more shallow.
I need to spend some time evaluating what would make the most sense: in the end Pathfinder doesn't really have an official subclass designation, it's just about what features I let the user select on the left side and have text in the main frame change based on (as opposed to just showing all options in the main frame). I think it's useful to show these selections to make it clear what choices players have to make apart from their class-independent selections like ancestry and attributes. But this is very much "down the rabbit hole" that I should look into later, I have way more important features to be adding. So for now I just define each class's selections as subclass or not in a config file (classes are rare enough that I'm not worried about this scaling, I might have to do 4 a year, and I can set reasonable defaults too), and am fixing smaller bugs.
Luckily, I've found an in-hindsight-obvious fix for my issue of needing "clean" pf2etools data to put Parsefinder on: we're running Git, so just do it on a different branch! Switch PF2etools to prod, hard-reset to match dev, run a Parsefinder run and copy the data to the repo, force-push the whole thing up, and now Cloudflare's serving Parsefinder's best on dev's latest, but pf2etools's dev branch still has "clean" data I can reset to for a future run. I'll eventually need to do an intermediate branch with a preview to make sure everything doesn't break, but for now only I have access to it, and I'm not sure this will ever be widespread enough to matter.
I've been submitting a lot of bug reports to Nethys as well as a result of this, which hopefully makes up for using their data as a source. They're making some more decisions that I personally find ridiculous, so once again it's a great time to be "blackjack and hookers"-ing it.
wraelk
4d ago
I find driving tremendously stressful, because I am acutely aware that unless I pay perfect attention to all the things going on in the road and respond accordingly, I risk death for me and my passengers. So from that perspective I could see road rage, your life is at stake constantly and the object of your ire presumably threatened it.
But I recognize this view is abnormal: driving is statistically one of the most dangerous things you do, but I don't think most people know that, much less internalize it. I probably don't internalize it properly either, I'm sure I'm doing things that are statistically risky while driving.
I wonder if this is an early precursor to the Internet's anonymity and echo-chamber issues? When someone's in a car they don't interact face to face with the other drivers (you can sometimes see a face if you strain, but it will just be a second or two snapshot before your attention has to go elsewhere, the equivalent of a profile picture), they're just faceless others that only come to the driver's attention when they do something stupid. Everyone in the car with you is somewhat on your side, or at least will rarely take another random driver's side or argue with you about the situation. And when you finally come face to face with the other guy, you've had ages to convince yourself he's evil and you're past the point of humanizing him.
wraelk
8d ago
In fact, I would guess it won't be long before it's easy to create AI-generated "reenactments" of culture-war sensitive events.
Somewhat relevant would be the Gabriel Paul Horcasitas case, where an AI video of the victim (Pelkey) was used.
I initially thought the video was in favor of lenience for Horcasitas, since it mentions forgiveness and the article talks about the sentence being tossed, but when I checked as I was making this post I realized I got it backwards: the video was created by Pelkey's family, who were pushing for the maximum possible punishment. The judge mentioned being moved by the video, and gave the maximum sentence. This was sufficient for the sentencing to be tossed on appeal, as the appeals court was not happy with AI influence. So not a problem in this particular case, since presumably resentencing will be uninfluenced by AI.
I'm curious if that position persists across time and situations: if a sympathetic-to-the-assailant family had made a video of the victim forgiving him and saying he shouldn't get a significant sentence and that somehow made it through the courtroom, would the sentence get overturned?
Parsefinder continues slowly: I'm in a really awkward position where I have the Nethys data and I have the PF2etools data, and I can merge them well, but I really need to get the first merge to be really high-quality, because my collision management strategy prioritizes either the PF2etools or the Nethys, and PF2etools generally has slightly better data where there aren't errors. But once I do the first merge, that merge would be PF2etools and I'd thus have to choose between overwriting other PF2etools data or not being able to fix whatever issues the first iteration of parsed Nethys data had. So I've had a bunch of false starts where I was almost ready to merge and suddenly noticed a major issue and had to roll everything back.
I've somewhat fixed that by improving my pipeline by getting PF2etools to run against an external data folder instead of its own, which means I don't have to put the whole 700 files in Git and then back them out, I can just generate them, run PF2etools against that external directory, and overwrite them with any fixes I need to make, only merging them to Git when I'm content with the outcome. But it's frustrating in that I can't really have an "improved, but not perfect" PF2etools just yet until I at least iron out all the big issues. Right now it's subclasses, which I totally missed the parser not capturing and now have to do a bunch of extra web crawl retrievals to get data for, as well as ironing out all the details of that. I have a feeling that's going to be a fair bit of this week. I also quietly lost my Syncthing pairing between laptop and desktop for the captured crawl when I changed some path details around, which means that alternating between the two computers has been dropping records.
In good news, I got ChatGPT to navigate me through setting up a Cloudflare Pages worker against the Github repo and my domain, and now I have the existing form of the site live at my custom domain, which is just a really neat thing to see. The site is structured in a way that'll let me host it there on the free tier, so all I have to do is finish fixing the data and I can start handing out access to my players. For now I've set up an email-code authentication gate with a whitelist of allowed emails: the site's entirely static with only local-storage user data, so I don't think it would really matter if it were open to the world, but I don't really want to find out I'm wrong about that, and even if I wasn't I'm a little concerned about it being a target for AI hate and/or "you stole our work" accusations from either of the two sites involved.
For similar reasons, I also moved to a new Github identity that didn't contain my real name and rewrote all my Git commits to have that identity instead. There's no upside to tying this to my real identity (if I wanted to use it on my resume this doesn't prevent me from doing it, it would just limit who ties those two identities together to jobs I'm applying for), and while I don't really expect to get witch-hunted, it's nice to have an extra level of abstraction.
wraelk
10d ago
I will note that it's possible for a game to have models that favor attractive-on-a-male features while not favoring attractive-on-a-female features, which would still be a technical failure while allowing this contrast.
My first thought here is Skyrim, where I've often struggled to make a properly attractive-on-a-female face without some very specific mods: Skyrim seems to really like wide, square-jawed faces that can work pretty well on a man but are rough on a woman.
At some point they had to make the call to not spend more time getting those features ironed out and to some extent that prioritization might be downstream of ideological goals, but it could also be a lot of other things that aren't connected to that.
I think this is totally possible, but I do also want to reinforce that if this was "we wanted her to look maximally attractive, but screwed up", it wasn't a decision that was fixed in Mass Effect 3. Miranda was redesigned, but still looks worse than Yvonne Strahovski: I'd argue worse than her ME2 iteration, though that might just be "coming even closer to the uncanny valley".
Many other characters get significant redesigns, most notably Ashley Williams, who I do find significantly more attractive in 3. So if it's a design failure in 2, it has to be paired with a second design failure in 3: absolutely still within the realm of possibility, but at least an unusual coincidence worth noting.
Huh, I assumed that was much better supported than it was. The Mass Effect Wiki provides a variety of pictures of her, that could reasonably be interpreted to my eyes as "tanned white woman" or "hispanic/part-black woman". My read was Hispanic, but I don't have good backing for that.
Reddit is as confused as I am (thread 1, thread 2) with a decent number of people suggesting she's hispanic, others suggesting part-black. One thing we see a number of is unsourced "the dev said she's hispanic".
I tried to trace that down and got this comment on the Mass Effect Wiki by the user Stormwaltz.
Also bear in mind that in the ME milieu, if you're wealthy enough, you can extensively customize the appearance of your children through genetic engineering. I was always surprised that no one noticed how much Ash looks like Salma Hayek - same coloration, same cheek/jaw structure. To me, that says that often people simply see what they expect to see (a caucasian), not what's actually there (a latina).
Does this random user matter? Well, Chris L'Etoile, one of the writers for Mass Effect 1 and 2 (but not 3) went by the name Stormwaltz in his prior work in Asheron's Call, with Reddit linking to a few of his archived unofficial Mass Effect dev diaries at the now-defunct stormwaltz.gamerdna.com. He also has a unique yellow color on the wiki, which might mean some special significance like being a dev, although I was unable to confirm that, as he seems to have removed most of his contributions.
I'll also note that her voice actress is Kimberly Brooks, a black woman. The original voice was E.G. Daily, a white woman, but she was replaced on the grounds that:
the developers did not feel the romantic chemistry they were aiming for between the character and Shepard.
(from Wikipedia)
It's possible that that's a HR-speak version of "we wanted to divest from whiteness" but I'm not going to just assume that, a voice actor change would be a pretty significant step and I think that would need some executive buy-in, so I'll take that at face value.
She also has a sister, Sarah Williams, who is a pale-skinned white woman, but she also looks nothing like Ashley (something that Reddit points out, a contrast that's also helpful to highlight how Ashley is not obviously white). Stormwaltz's comment suggests that this might be a choice of genetic engineering, and since we never see parents, we don't know if it's "darker-skinned parents engineering a white child", "white parents engineering a darker-skinned child", or "parents were some mix of skin colors and we got two extremes of the genetic range from un-customized children". It's an open question in the sense that it's heavily dependent on your choice of axioms if a genetically-engineered Hispanic-genetics child of two white parents would count as Hispanic.
So the best info I can get is fairly ambiguous. But if Stormwaltz actually is Chris L'Etoile, I think that's sufficient to support my point: if the writers think she's Hispanic, that's what will matter for decisions about who to uglify and who not to.
wraelk
11d ago
Would you consider Bayonetta (2009) or Rise of the Tomb Raider's Lara Croft (2015) stylized? I wouldn't, but maybe I'm not thinking of it the same way you are.
Just above someone pointed out Ashley Williams from Mass Effect 3 (2012) which is even the same series (and they didn't fix Miranda in 3, if anything they made her worse).
wraelk
11d ago
It's a fair point, but nevertheless I have to notice that Miranda Lawson, the genetically engineered perfect woman, is notably uglier than the actress they modeled her off of.
And they don't fix it in 3 with slightly better graphics technology, either: if anything she takes one step further towards Michael Jackson.
It could have just been a miss on design that they didn't have time to fix, but it would be interesting to learn if there were conflicting interests at BioWare at the time. The person designing the character probably isn't the person picking camera angles and animations for the cutscenes.
Ashley is a good counterexample, but it also raises the point that making attractive characters in the engine is possible, because they did a good job with her, so what happened with Miranda? I wonder if the fact that she's non-white changed things. Now granted, she's also one of the most racist characters in the series (not that that's that much, but it's enough for some pearl-clutching), but again, the character designer isn't writing the dialogue.
Publishers did not switch to the modern prudishness until about 2015-16, post-gamergate, and it was a sharp change.
But the reason Gamergate was so explosive is that a lot of those pieces were already in place: it wasn't that Gamergate was a starting point for progressives to make their way into video games, it was that there were so many there already that when the Zoe Quinn flashpoint set things off and made people reveal their true colors, the battle lines were instantly everywhere. Many of the people who would be fighting on Twitter in 2014 were working in the industry in the preceding years.
Publishers were setting direction, but lower-level workers were making decisions too.
Also, in some cases of uglification there is a female human actress who is very attractive, far more attractive than the character based on her.
One of the earliest examples of this I can find is Miranda Lawson from Mass Effect 2, who was based on Yvonne Strahovski and canonically was genetically engineered to be perfect, yet somehow looks significantly worse than Strahovski. It's close, so the most charitable case I can make is that they started with a picture of Strahovski and changed it without recognizing that those changes made her worse, and/or engine/artificial lighting constraints magnified the issue. But Strahovski isn't perfect (her face is oddly wide, for one) and they certainly could have worked to improve it rather than the end result they got.
Less charitably, BioWare in 2010 was probably decently progressive: no M-M romances until ME3 (some F-F, but that's as much fanservice as it is catering to progressive beliefs), but Jacob was apparently planned to be an M-M option in ME2, so this may have been an early "can't make her perfect for the male gaze" despite her struggles with being designed to be exactly that serving as one of her major character arcs.
ME2 was debatably the last good game BioWare ever created, so that may have been some of the cracks showing before the steep descent to the depths of Veilguard.
wraelk
11d ago
It also has the very sizeable drawback of people regarding the opposite gender as other, as categorically beyond one's ingroup; I think that might be what Dr Roth is getting at.
I wonder if this is just reinventing chaperones: we want men and women to be interacting so they get to learn about each other, but when left alone there's potential for abuse. Thus, allow intersex mixing, but only in controlled spaces with authority figures that ensure safety.
You'll get people sneaking out as a second-order effect, but this both shows a high level of consent on the part of both participants (since they had to defy the official rules to do it) and adds some level of "well you hopped over the safety railing, what did you expect?" It also self-corrects in that the answer to "this bad thing happened to this person" is "stay with the chaperones".
Of course, that fence got torn down, so presumably going back to it would just lead to the same sequence occurring again.
wraelk
11d ago
Making it so anyone that doesn't have this level of informational hygiene can't sell drugs seems good on the margins. There's nothing stopping this approach from being done now, and yet people are still arrested as drug dealers. The higher you set the lower bound on intelligence for drug dealing, the fewer drug dealers you get, especially since higher intelligence gives better alternatives for a career path. They'll need intelligent mules as well, or the police could just keep an eye on locations they've used before.
This particular approach seems vulnerable to scams, as well: just spread your prepaid number as a "drug ordering number", take the money, then direct the buyer to a random location (where there aren't any drugs hidden). When they complain, claim that you put them there and someone must have taken them. Word will probably eventually spread and people will stop using it, but you can just create another one.
Now you need a second communication system to legitimize the numbers, and you have a problem: if the communication is anonymous, it's easy to spread false information leading ultimately to scam numbers. If it's not anonymous, initial information about numbers can potentially be traced back and used to identify drug dealers.
Actually, what's stopping assorted street junkies from just identifying the drug distributors, following them to the dropoff, and taking their doses before the customer does? There's a limit to how much force the anonymous dealers can exert, because if they're regularly sending thugs to beat up thieves, now the thugs are on police radar and can be traced back. Maybe some sort of thug-for-hire service where they just get an anonymous telegram text with a picture of who to beat up, but there's still nothing stopping the police from arresting them.
wraelk
15d ago
The way you win the news cycle is by talking about the things YOU want to talk about, not the strange and unexplained disappearance of your own political candidate.
This is the strategy I'm proposing, I think.
You find your candidate story, throw one or two news articles out about it, deliberately structured in a way to be suspicious, then say nothing more about it (it's not a story, so why would you?), and let the conspiracy-hungry side of the Republicans pick it up and waste a bunch of time talking about it. You don't just keep throwing articles at it the way they did with the Haitians, you largely ignore it. If anyone asks why you're not talking with it in a situation where it makes sense to say something, the line is "nothing happened so there's nothing to talk about". In the meantime, you just talk about whatever you were otherwise going to talk about.
Then later on you just toss out some rock-solid proof: you don't spend a bunch of time talking about it there either, but you make sure there's ironclad information making it very clear that nothing's going on (not in a "there are stories debunking this" sense, in a "anyone looking at this evidence for more than a minute realizes there's nothing going on" sense) and it's easily accessible.
The examples you propose are mostly the media trying to force an issue by throwing out a large amount of news volume, and indeed a lot of it backfired, both in that it got people thinking about it more than they otherwise would have and in that it caused the media to waste airtime they might have been able to use somewhere else. I'm proposing sort of the opposite: spend little time on it in terms of news volume, but bait a hook for Republican conspiracy theorists to bite so that they waste their time and credibility on a story that was poisoned from the start.
wraelk
15d ago
The NYT claims to have seen home video proving that. You don't get to, though. Nothing suspicious about that.
It's occurred to me before that a potentially valuable tactic for a Democrat-friendly news media would be to take a complete nothingburger of a story about a Democratic candidate and just act really weird about it. Assure everyone that it's nothing, but make oddly specific wording choices, show less evidence than you have on the grounds that it's not all that interesting a story. Try to get Republicans to bite and waste a lot of time and effort hunting down a story that goes nowhere instead of stuff that can hurt the candidate, maybe fracture the base a little fighting over it. Then make sure you release the completely uncontroversial evidence that proves it's nothing in plenty of time for it to disseminate before the election, knowing some Republicans are dug in at this point and will keep wasting airtime trying to argue a conspiracy around nothing.
The whole Obama birther thing might have been an example of that, I'm not sure how deliberate it is, but it seems like it would be if anything more effective today.
wraelk
16d ago
I really appreciate you revealing the twist, because up until now I had only seen media that didn't mention that point and had been very confused about how this could possibly be seen as good.
I had to pause the trailer halfway because it felt like it was an AI-generated trailer trying to make a joke I didn't get, and only resumed when I was sure it was a real movie.
So the reveal that this is intended is somewhat of a relief, in that it's "Hollywood is trying to do something clever that didn't really work" as opposed to "Hollywood actually thought this was good".
But ultimately this is just "I was only pretending to be retarded", where I'm not sure why it should matter to me the moviegoer that they wanted to make a bad movie. And I have to imagine some of the reason it's bombing is that a lot of people are not going to look past the base impression.
Which is funny because it speaks to some myopia on the part of Hollywood. If the public is giving movie creators some latitude and going "well no, this can't actually be that bad" they'll at least get to the main point, which presumably the creators thought was good. But that's probably not how the public is viewing new movie releases these days, and as a result it lives or dies on the perception of the thing it wanted to be deliberately bad. And unless someone spends a bunch of time looking into it later, it will just be another point supporting that Hollywood is cranking out loads of terrible movies.
I'm working on a highly vibe-coded project that I've wanted for a while.
I'm an active GM for the D&D spinoff Pathfinder 2nd Edition, and one of the selling points is that all the mechanical components are available online for free: every class, monster, feat, item, etc are freely available on a site called Archives of Nethys as soon as the volunteers get to it. This is nice for budget-minded players, but is also great for researching builds and such since you don't need to worry about paywalls gating your ability to check the content.
Unfortunately, I'm a little disappointed with the structure of the site: the pages are bulky and while they have a very nice search system, it still doesn't feel easy to browse. There was a much nicer site, PF2etools, that is much more responsive and easy to navigate and compare. But they stopped updating the site years ago, so it's wildly out of date with the latest books. But the bones are good, the site is entirely open-source, and it just needs properly-formatted JSON to run.
So my project scrapes HTML from Archives of Nethys (moving at a glacial pace as I really don't want to make this a whole discussion with the Nethys volunteers, especially given the wildly AI-hostile moods in geek-related habits these days), parses it to PF2etools-formatted JSON, and merges it into the existing PF2etools data corpus, with the intent that I can bring the codebase up to date with all the books, and potentially rerun it periodically to stay up to date. Since it's a relatively small static site, I'm hoping I can just host it myself, possibly behind some authentication check so I'm not having discussions with the Internet about this, and have an up-to-date copy for myself and my players.
Long-term, I've got some other plans:
- Integrating errata overlays: every GM has to make some house rules, and I'd love to have them integrated right into the text of the site so anyone can quickly look up "how it works in my game". This would ideally allow me to do bans and homebrew at the same level.
- Supporting more formats: we don't really need Nethys, the dream would be just pulling directly from a book PDF (which would also let me grab third-party stuff that wouldn't be in Nethys). We could also potentially output in other formats like Foundry-compatible JSON, easily-readable text, or anything else that might be useful.
- More integrated tagging: PF2etools has a few unique-to-it spell tags like "forced movement" that make a lot of sense to group and tag game pieces with even if they don't have any official game designation. I'd like to be able to standardize this process and add a lot more.
- Ratings support: being able to load a content creator's evaluation of something like spells and have it mark up and filter spells based on that.
- Full Scryfall-style query syntax
- Going really crazy and trying to build a full character builder off of the existing JSON database. The hard part is getting all the options encoded properly and I've done that, so could I just build something to compete with Pathbuilder?
Currently, I've got a full crawl of Nethys that's translating into PF2etools-compatible JSON and merging properly into the PF2etools corpus, but I keep running into minor fixes when investigating that make me roll back the whole merge: the latest two are that it was treating heritages for any ancestry as their own ancestry, and that Nethys randomly decided to treat individual chapters of some of the latest books as their own sources (a change that is making me more and more confident that making my own database with blackjack and hookers is the right choice).
It's getting very frustrating to try to check these things. I'm floating the idea of trying to get ChatGPT to build some kind of review queue for me where it points out things that might look unusual or wonky and I can see if it looks the way I'd expect it to or if it might point to some error in the parsing progress.
wraelk
18d ago
It's possible she's trying to make that case, but I think it's more that she's presenting magic as such a huge force multiplier that you can just have a very small number of people ruling over the masses. While teamwork certainly helped against Voldemort, in the end it was mostly just about out-magicking him, not "the people rising up and overthrowing tyranny".
Voldemort's plan seemed to be "I can kill anything you can organize against me" and while that might not have worked against an army, the wizarding world doesn't seem capable of fielding one.
wraelk
18d ago
I'm not questioning the effectiveness of this at all, but there's something inherently funny about "clean your cutting board with a laxative", in the sense that I'd expect it to be a key prank in a movie about rowdy college students sabotaging the dean's party.
Dean Nofun, after watching his guests have to rush to the bathroom after every course, perplexed, begins sniffing the cutting boards with increasing concern... "FUNLOVING HOUSE!!!!!"
wraelk
20d ago
I don't think AI can get me arbitrarily good music in any genre, but it's quite good at producing music I like to listen to in bulk.
My programming/work standard now is Synth Mage, a synth metal AI channel that puts out music in hourlong blocks (the second hour on the video is just a repeat). I got ChatGPT to build me an ffmpeg script to split it into individual tracks based on the description timestamps (which unfortunately weren't accurate for videos where the author put an intro, but I'll get around to fixing that eventually), and now I've got a multi-day music library of stuff I could listen to all day.
The big thing for me is to avoid lyrics: AI lyrics still seem to be a little weak without a lot of human guidance, though a couple of Synth Mage tracks do just kind of treat voice as another instrument and that seems to work pretty well.
wraelk
20d ago
This is a great point I forgot about, probably because we don't really see much of those groups or get a feel for how big they are. Harnessing the oppression of groups that the current government doesn't favor is a lot more like a populist movement than how he was interacting with the other wizards.
I guess the monstrous groups figured any port in a storm: infighting in the wizard government would give them more freedom to do whatever they wanted, and even if Voldemort wasn't interested in keeping them around, he's got a much more ambitious project to work on in conquering both worlds, that might give them time to get out from under the wizarding boot and be ready for when he finally comes for them.
If they want to decide who gets dates, they're not going to do that. I'm skeptical that a general public who would otherwise be interested in the app would balk at it being closed-source.
The mechanism probably looks more like "when you're a disfavored group, the partners you want to date and who would be interested in going on a date with you are on the app, but the app won't match you with them". So favored groups get easy access to their desired partners, disfavored groups have to try and use other channels. There's a limit to how much this will work if the desirable partners aren't happy with their dates, but there's probably a fair bit of thumb that can be put on the scale.
More options
Context Copy link