gattsuru
6d ago
The notices bulge one is a furry meme, rather than trans. (Not even girls-with-dicks side of furry, afaik; I've seen it more from the gay side, and not just in the sense that I would see more of the gay side.) Kinda has escaped containment since it originated as a bit of an anti-furry thing making (fair) mockery of cringy RP conventions, so might just be general too-online reference.
gattsuru
6d ago
The alleged shooter has been caught, and this time they seem confident enough about it to give a press conference.
We've gotten more details about the etched messages on shell casings, and they seem a mix of general antifa and memelord. The supposedly trans-related one looks to have actually been the furry-originated 'owo notices your bulge' joke, and that's usually more cis gay (or even, rarely, cis M/F) roleplay before it turned into a general cringey joke, but along with the rest of the details, it's looking pretty explicitly like radicalized iso-standard lefty.
I can't link to it without self-doxxing, but a teacher in a nearby school, one who I worked with for STEM outreach post-COVID, is now in state-level news for the same stuff as the Home Depot lady from months ago. I'm pointedly getting out of any chat or social media around that entire conversation and its downstream effects, and it's too early to drink, and I'm going to be working with students from that school later today and I'm not going to mention it and I used to respect him and I'm done.
gattsuru
6d ago
The 'she wasn't a virgin' defense to rape allegations didn't get accepted by juries very often in the modern era, but it was still pretty costly for genuine victims even where they were successful in getting their attackers brought down. What extent people don't bring charges in marginal cases where those costs are high is an unknowable number, but it's probably not zero.
gattsuru
6d ago
I'd hope he's learned a lesson or at least fear, and you're probably better at reading body language than I am, but Hasan specifically also has a pretty long history of pretty explicit calls to violence (cw: BakedAlaska link, also BA-grade legal analysis), and his career depends pretty heavily on not having a legal hammer or Senate nudge go after him. I have no idea what Twitch executive he's got compromising photos of, but given that the website is pretty notorious for banning game streamers who get too rough, there's gonna be some fallout on that sphere.
gattsuru
7d ago
What do you want to say at work that you think you're being prevented from saying because of potential employer liability under "hostile work environment" standards?
It's a matter of standing law that the Civil Rights Act controls what public radio stations that employees may turn on. Google defended -- and the NLRB accepted -- that anti-discrimination law actively required that the company police the speech of its employees. Other cases have held that employers are responsible for even off-premises and after-hours speech by their employees, or where the speech was not even directed from one employee to another.
What makes you think than your employer would have no problem with you saying that even if the potential liability didn't exist?
In some cases, this is plausible as a defense. Several early hostile work environment cases revolved around 'employees' who were already fired by their employer, with the lawsuit between plaintiff and employer revolving around whether the employer should have acted sooner.
In other cases, it's hard to even separate matters; there's now a strong convention against nudie mags in even the bluest-collar of blue collar jobs, and of course no employer wants their workers to be staring at breasts while on the job today. Would that have been considered as unacceptable without thirty years of HR hammering into every employer and employee?
But in most, it's not especially defensible. We just had a big court case about an employer making fun of an employee for being gay (and fat); the only reason the employer won (after a long and uncertain court case) was because everyone agreed this job was ministerial, not Because Free Speech Uber Allies. And the employer very clearly did not want to apply the anti-'hostile work environment' policy, given that they probably spent tens or thousands of dollars defending their not complying with it. There are hundreds of cases like this, almost all of them get no legal defense, and that's before getting to the wide variety that no one defies even when they want to because they know they're fucked.
And then you think for ten seconds, and you remember that people put a ton of political capital into not only maintaining but expanding (Bostock! Kinda a big deal!) these policies, and it becomes kinda obvious.
gattsuru
7d ago
This is a fascinating normative statement, and one I'd love to support.
As soon as we turn from 'should' to 'does', though, the answer changes radically. Mike Adams was forced into early retirement (and driven to suicide) over his personal writings in 2020. Damore doesn't have his old job at Google back, and the punchline to his whole NLRB thing was Google arguing (and the board accepting) that the law required them to fire employees for speech. People were fired for anonymous donations to Kyle Rittenhouse's defense fund. Nor does it stop at firing: Kyle Kashuv and Harvard, LexManos and Forge, Vaxry and Hypr, Mercedes Lackey and the convention circuit, yada yada.
There was a big important court case about whether the federal government can pressure private companies to ban and censor specific users, and SCOTUS said fine by us. [context]
Never again would be a wonderful philosophy. It also demands that it stop happening the first time. I would love to see that change. But I notice that it is only when progressives are getting fired that any progressive cares about freeze peach, even the ones that proclaim they were 'always' the principled ones.
I would love to have arguments against this strategy; I don't.
gattsuru
7d ago
Punchline to the Pause Letter story: data was 'lost'. Indeed, FOIA-able records for the relevant periods in question here were 'inadvertently lost' years ago, and we have only found out this month. Even after the FOIA was submitted, the SEC did not initially search these records, even after claiming both orally and in court filings to have completed and re-completed a search of all records, and unsurprisingly no disclosures of possible record destruction were ever delivered despite explicit regulatory requirements -- even long after the internal report about the scope of the 'inadvertent' data loss was complete.
Most critically, we did not get this by the grace of the courts.
gattsuru
7d ago
The issues show up without diving into algorithmic media.
gattsuru
7d ago
Probably a reference to this conversation and its predecessors.
gattsuru
7d ago
gattsuru
7d ago
Most of the coverage is downstream of this WSJ article, which is unfortunately paywalled in a way that archive.is can't bypass; see here for visible-text.
Given how poorly that Red Hood comic went, I'm not sure Felker-Martin's barely a tenth of a Tara Strong. Dowd is more persuasive, with the caveat that it falters if pulls a Toobin and is back in six months.
I guess this is where I should clear my throat and say "Political violence is bad and I condemn the killing of Charlie Kirk"?
I genuinely do appreciate that. I will note others: I've mentioned KelseyTUOC already, but there's been some number of decent prominent and not-so-prominent people who've spoken up, sometimes even in credible or costly ways. Even some pretty awful scumbags are at least trying to motion around it, if not very sincerely. Some of them are even sincere-seeming: I genuinely neither expected nor hoped a Young Turk to have tried, even if he's still pointing the wrong direction.
It's also a long way from persuasive enough. This is a moderator at NeoForge Discord. This is a moderator at the Hexcasting Discord (to her credit, the mod dev herself has been more responsible, albeit in a 'don't make people watch someone die' sense than a 'aggressive violence is bad even when it happens to people I don't like' one). This is imgur yesterday, this is the front page sorted by viral today. This was tumblr yesterday, hastag his name, sorted by top; this is tumblr with the same constraints today. I logged into Star Citizen last night to take my mind off things, and had literally could not get out of the in-game bed before I had chat cheering it on; this is from an FFXIV guild I dropped before the election, and a discord I'm gonna leave in a few months.
And it's not just the nameless and faceless grunts, or bluesky, or the people who skinsuited a project I once respected. This is Ken White, who to what minimal credit he deserves says that Violence Is Bad before going straight into 'you can't defame the dead' mode. This is Barry Deustch, B from Radicalizing the Romanceless. This is from the writer of NeoReaction: A Basilisk, and was well-respected in the tumblr ratsphere for almost a decade. There were 51 posts over 12 hours in the rpg.net thread (cw: big image), and while there's a couple that aren't dancing in blood, there's literally five times as many where people who I once took seriously now going full :
On the other hand, my immediate reaction is fuck them, they get NOTHING.
It's like the person upthread saying how we shouldn't have this thread at all - is this how we defend actual free speech and small-d democratic values? By running and hiding and staying quiet? Nah, screw that. Charlie Kirk was an awful person and I feel bad for the family he left behind, but I would feel bad for them beforehand, too, for having that kind of guy as a father and husband.
Now, anyone getting revenge porny or actually cheering on political assassinations isn't correct, either. Frankly, given how parts of his fanbase were turning on him from going from 'release the Epstein files" to "it's a Democrat hoax", I expect this to be part of the occasionally seen far-right "you aren't hating enough or the right people" circular firing squad, which unlike the leftist version tends to involve actual firearms.
I can keep doing this, if you'd like, but I don't think it's healthy for either of us.
I've gotten it from someone I let live in my home for six months while they were getting back on their feet. Didn't even go looking for them, I don't follow them on tumblr anymore, just bam, snuff video with a Dark Souls meme thrown into it, with a 'leopards eating faces' tag in case they needed to make it clearer what they were condoning. Do you want a list of exactly what thoughts, in what order, went through my mind? It's not just me; KendricTonn is another guy who fled to the heartland (poor bastard ended up in Ohio!) and he's getting it, too.
I considered looking up the social media of some of my past partners. Do you think it would help, or not? I'm pointedly not doing it, because that way lies even more psychosis than looking up your exes normally does.
As a prediction, which I will send you in plaintext in PM and post publicly here in a week, sha256: a009fcb948bd1a70a38d133d81f0cc96af6efa94904133184a5f40d0cb5d6004
Because ten years ago, it would have been useful to have a decent handful of examples of prominent speakers who would consistently speak in defense of "bad argument gets argument, not bullet". We have not been dumped ten years in the past. We have a decade of people Friedersdorfing these grand principles about how they'll defend people that they totally didn't defend in the past.
Jerk That Can't Write A Comic Story Worth Shit isn't a costly signal. Matt Dowd might be, if it sticks. Actually blackballing people and organizations that promote or defend this sorta stuff is. Either people haven't brought serious and costly signals of enforcement against their own side, or people think these examples you're bringing forward are the serious and costly signals. If that's the central example from the aftermath, I'm going to point to Forge again, and Damore again, and Kashur again, and show exactly how much political debt their alliance is in.
They might not have done it, themselves! They might even, in their heart of hearts, have whispered words about how it tots would have been better if no one did these terrible things. It's genuinely terrible that people have to handle the weight of bad acts from people they might not even like, just because their political alliance. It's also a little late for them to complain.
gattsuru
7d ago
Fair. We'll see if it's a real thing or just another Toobin, but I'll register that I did genuinely not expect that.
gattsuru
8d ago
What would that even look like, with tribal positions reversed?
The simple answer is a journalist, and to be fair that has happened, albeit typically with different political valences and, probably more critically, at much more personal scales. And there's a fair cynical analysis that scale matters, and it shouldn't.
The more morbid answer is that it's literally anyone with a social media account. We had discussions here where posters -- well-respected ones! -- thought some level of political engagement put people beyond questions of doxxing or targeting, and pointed to someone who could e-mail and call state politicians, and motioned about how great power came with great responsibility. Before today, you could imagine a world where that had limits, not just ones that the person drew for themselves, but that applied across their political alliance, and 'responsibility' meant nothing more and nothing less than being set right when wrong. I can hope that's what the posters here meant, back then.
Even if we turn down the violence, and I'm not certain that's even possible, at the end of the day there's no going back to that. This is not someone getting kicked out of college, and it's only deniable in the sense that people will deny it.
Because there are countless ways to say Kirk wasn't harmless; my tumblr and bluesky and discord feed has no small number of them, and I can't even log into an MMO without seeing it (thank you Star Citizen's chat being broken like everyfuckingthingelse). He was impactful, and it doesn't take much impact at a national level to change tens of thousands or hundreds of thousands of lives, and even if you think it was for the better not everyone's going to agree with you. And then you remember Joe The Plumber or Ken Bone, and weep.
We've seen variants of this before -- the doxxing tit-for-tat back in the CCW days cut my teeth and a few others -- but they depended on certain very specific tools, which could be shut off. Now, the only necessary tool is a network account.
gattsuru
8d ago
Outside of the general ethical problem with 'government agents did bad things, these guys are also government agents' -- it's hard to overstate how indiscriminate -- McVeigh specifically evaluated the daycare situation in the building he targeted. That's a good part of the point of the original charcoal briquettes rant.
gattsuru
8d ago
Yes. I dropped a PM conversation with FCfromSSC over a tangentally related matter, and there were a portion of discussion around the Baude/Paulsen stuff that I will not post here and did not write down anywhere except in a direct e-mail to Baude.
There are ideas that are dangerous. I will no sooner than write publicly how to kill large numbers of people and probably get away with it than I would inform twelve year olds about the cool cleaning technique of mixing bleach and ammonia.
gattsuru
8d ago
gattsuru
8d ago
I would ask that you think, for about ten minutes, about what those dominated fields will look like in a sustained conflict where a) many of the foot soldiers will have trained themselves to enjoy their 'enemies' pain, b) where the 'management' tier is primarily interested in farming those foot soldiers by making alternatives to violence unimaginable, c) where a man who killed someone in a protest over a year ago is walking the streets on bail, today, d) those 'management' personnel will optimize for the survivability onion.
gattsuru
8d ago
There's a bit easier of a way to square the circle: even trivial consideration can give a pretty sizable laundry list of ways to kill someone from a long range, very messily, especially when half of your compatriots agree with you. Some of them are even easier to get away with!
Further information not available here.
There's some funny nutpicking, and then there's some not-so-funny nuts someone else picked.
I'm... hard-pressed to comment more, here. I'd like to 24-hour-rule this, because early reporting is so often so bad for these type of incidents, but there's almost zero chance it's some nutty right-winger who thought Kirk Didn't Stalin Hitler Enough, and at most it's a question of whether early reports that it's trans-related-stuff are right, and my gut makes it pretty hard to think they're not or that it'd even matter much if they weren't. It's a little nice to see some lefties or 'centrist' liberal saying this is bad (woo, Kelseytuoc retweeted aghamilton28), but when that's not even close to universal, and when those who don't or who find it worth other forms of comment don't get at least a shush, it's hard to take too much solace from the exceptions.
But it's soon, and maybe in a week there'll be some mirror to the culture war WMDs of the past and I'll eat crow and be very happy about it. Or maybe some manager will steamroll through MSNBC and The View. But I'm not optimistic, and I dunno if we'll even get Home Depot Lady v2, and I don't think anyone on the conservative side of aisle will or can be persuaded to care about that.
I've been trying to write a followup on the Paul Kessler thing (trial: supposedly next month, but also supposedly two months ago, maybe early next year?), a contrast on a lot of other public violence cases in the aftermath of various public protests, and on a bunch of recent gun cases, and there's some increasingly obvious answers for a lot of it. Four years ago I worried about things going hot in a crime-of-passion sense, where plausibly-reasonable decisions run into foreseeable consequences, and the ramifications spiral and reinforce each other as 'our' definitions of reasonable decisions increasingly disagree, with a standard example being where CPS and trans kids and the horrors of the foster system run face-first into each other. The answer seems, increasingly, that those crimes of passion will happen, and the people charged with enforcing the law against them will decide to bring the hammer down or not based on their current wins, and those whims have unsurprisingly kept pots from boiling over, and that's a solution of a sort, but only until you think about what comes next.
There's a lot of discussion around this topic that I just aren't willing to publish publicly, because at some point, someone that isn't a garbage person is going to take ten hours and seriously think about what the results and ramifications and incentives are, and how they can maximize their impact. And it's going to result in hundreds of deaths, if not thousand+, in the first incident, and we won't find them, and people won't even wonder why so much as who they can blame.
gattsuru
9d ago
For imagegen, you've got two generalist options:
- Automatic1111 (install guide is the moderate tech-savvy option. It's relatively easy to set up (ie, will handle your python venv and download a default model for you, though I recommend looking for more specialized models from civitai yourself after setup), and while it exposes some complicated options, you can just start with the prompt and negative prompt and get some outputs first, upscale them, and/or do basic img2img. There are some powerful options built-in, and even more with well-supported plugins, but you're not going to have to go into hackerman mode to get anything out of it.
- ComfyUI (install guide) is the more complicated and powerful one, at the cost of being a little (more) obnoxious to work with. To do anything, you need to set up a workflow made of multiple nodes, and while they're easy to pull a simple template or to download prebuilt workflows, it can be a little overwhelming and it's always a little obnoxious to get used to. It has much wider support, and support for other types of models (eg, with plugins Wan V2V generation), and can even queue a bunch of wildly different imagegens (eg, queuing different models, settings, or even workflows), but at the cost of taking a lot more time to handle.
There are some specialty cases (eg, Wan2GP is like Automatic1111, but only for running video models on mere-mortal-level GPUs; a big stack of options for 3d model generation), but those are the big ones.
For classification and categorization, there's a lot of options, but most of them are intended to run on servers with less powerful graphics cards passively, rather than on-demand from a desktop client. The three I've tried are PhotoPrism, Immich, NextCloud Memories. All worked well enough for my purposes, but the user experience and setup difficulty is wildly different from one to the other -- I'd probably point to immich if you are okay with Docker now, and NCM if you absolutely won't, but there's a bunch of tradeoffs to each.
I'd assume there's some desktop tools for this, but I haven't found any that were good and turnkey. You do have the VRAM necessary to train your own AI classifier (I'd recommend YOLOV4 using WANB) pretty quickly if you've got the training data, but it does take a lot of preclassified photos to train it (>200 per category minimum, imo), and you'll need to do some (high-school-level, simple CSV munging) code to actually do the sorting or tagging.
gattsuru
10d ago
Grok's... weird, performance-wise, especially compared to ChatGPT. It (and especially 4) are heavy thinkers, and from local use with Qwen I'd expect it to have a bigger beneficial impact, but there's something weird going on with whatever RAG-like they're using that makes it go wonky at times.
It's very prone to math errors, even by the low standards of LLMs, both 3 (free) and 4 (paid) still hallucinate and gaslight pretty badly, especially when you get even slightly off the beaten path (compare this to this on the IMU implementation: Viture does actually hide their documentation in a bad pdf, so it's not surprising that neither could find it, but it is disappointing that Grok hallucinates a non-existent class).
On the other hand, Grok's been surprisingly good at handling 'real'-world questions, where others at best seem to give okay answers. There's a risk of landmines because it does still hallucinate and it's (ime) more persuasive than ChatGPT or Claude when it does, and it doesn't stop you when you ask a stupid question (ie, dollar/GB is a really misleading metric for almost all use cases), but my experience is that you need to keep that in mind for pretty much every LLM. It's been a much stronger tool for helping teach, if a bit verbose, even for coding questions.
I dunno if I'll stick with it, and especially for coding-focused use cases I can't really recommend it. But if you don't mind how verbose it gets, there are some types of questions it does better.
gattsuru
10d ago
It depends on the exact implementation, but most frontends have to provide the LLM all or a large subset of the previous conversation as an input for a conversation to meaningfully continue. Where context windows are small, they'll have to truncate early portions, use summarization tricks, or use tricks like rag. Even using those techniques, or for LLMs with very long context windows, an LLM given both 'forks' as an input will usually seem very incoherent very quickly, as it will put information, requests, or status from the 'other' branch -- even the best-case scenario would be much more similar to asking the LLM option A and then option B in sequence, rather than separate branching options.
That said, even LMStudio supports just branching a conversation with a single click. I think you can technically do it with ChatGPT/Grok by abusing the Share function and just using that linked conversation as a separate branch, though it's a little more annoying.
gattsuru
13d ago
Trump will have one of his more shameless toadies in Congress (looking especially at Andy Ogles, who’s already happily floating this) propose some kind of legislation to allow Trump specifically to serve additional terms. If the GOP manages to retain any kind of majority, I expect that legislation to get put to a vote, and ultimately pass along party lines despite initial token unease from more ‘moderate’ Republicans.
I'm curious what you'd be willing to bet on the condition that such a bill never makes it to the House and Senate floor for a final vote, conditional on Trump surviving (and not being comatose, Biden-esque, yada) to January 2029? Because at 80% confidence... well, I won't call it free money, but I'd be willing to bet at greater odds the opposite direction.
First the violent nutjobs took FLCL references, and now they're taking the concept of getting rawwed by a werewolf in tight jeans. Is nothing sacred.
More options
Context Copy link