#data lab

[catboybiologist]

By now, there's lots of people have heard about the internal CDC memos for all newly prepared manuscripts (like future scientific papers waiting to be published):

There's so much to comment on, and I'm seeing it all right now. What the state of science is. What this means for the queer community. All of that.

But fuck, I think I might genuinely start crying over this. As a transgender biologist, this feels like a brutally personal blow. I slowly accepted my gender alongside my biology education. The more misinformation that was spewed about "biological sex" by mainstream media, the more my professors, colleagues, and primary sources would casually drop information that proved they have no idea what they're talking about. I'm not an expert on sex determination, gender, or transgender biology specifically by any means. But my worldview has been crafted by my studies in genetics and molecular biology.

Engaging with this research helped me demystify transition. It helped me optimize my transition. It helped me explain how HRT and other steps of trans healthcare work to other people. And it helped me overcome my own internalized transphobia, and finally start transitioning, despite knowing I wanted to since my preteen years.

Who knows how enforceable internal guidelines like this will be. But its certainly going to scare a lot of researchers away from transgender healthcare and science in the coming years, and that breaks my heart.

There's a lot I can say here, but fuck. I just needed to vent for a moment. Fuck.

[lorenzonuti]

Whispering secret data.

[stars-obsession-pit]

Danny had been put through serious hell by some ghost scientists from the GIW in the name of “research”, and his ghost form remained incredibly unstable in the aftermath. Even during the stretches when he could best hold himself together, he was still visibly covered in wounds.

Clockwork did offer him the option of being de-aged to recover more smoothly, but he’d said hell no. He had absolutely zero desire to be further reduced—especially not mentally when his mind had been one of the few things the scientists couldn’t take from him.

Besides, he has a job to do.

Though he may have escaped, not everyone else had. He refuses to leave them in that nightmare.

And when the news caught wind of a his attacks, it exploded into a massive media circus.

No matter how much the GIW tries to spin Phantom as an insane monster, a heavily mutilated teen attacking secretive labs does not paint them in a good light.

[warmfuzzyanimal]

new icon, directly inspired by my morning walks ^__^ 🌲🌲🌲☁️

[applepixls]

not quite sure where this originally comes from but i saw this picture and couldn't help but think of the permit office people-

[crow-caller]

A lot of people think autism research is solely this vague grouping of evil non autistic people guessing at things from afar, but as much as that happens, I want to inform you with insider knowledge a lot of modern autism research is done by autistic people!

And a fun fact related to this: "autism" is a common special interest! As in, a lot of autistic people have autism itself as a special interest (esp women, perhaps bc they're likely to be late or self diagnosed). People with a special interest in autism are also more likely to get involved in autism research as participants, and thus there's a known overrepresention of it as a special interest in data

[waitineedaname]

I've finally done what I inevitably do for nearly every fandom i get into: I've made a deranged spreadsheet tracking some linguistic element in a piece of media! this time it's tracking the usage of swears in Scum Villain. if you're curious, here's the full spreadsheet, but I'll write up my findings here

some notes on my methods!

I differentiated between internal and external use. external use is pretty straight forward, that's just dialogue said out loud. for internal use, that's when it's used in the narration. obviously, this is skewed towards Shen Qingqiu as our primary POV character, but occasionally the narration dips its toes into other characters' perspectives, so I kept that in mind and assigned the usage to whoever seemed most relevant in that snippet of narration. also, in the case of the transmigrators, I also counted speech directed at the System as internal because they're less likely to self-censor when talking to the System because no one else can hear them

this was done by searching for specific words and their variants in online copies of the official English translations of all four books. there are a bajillion different swears in the English language, but I tried to focus on the most common/most likely to be used. also, since I was searching for the words manually, there are going to be some that are missed because they're written out abnormally, like lengthening the word for emphasis (eg. fuuuuuuuuuck) or censoring it (eg. b- [BLEEP]). side note, i don't know why the hell SQQ ever censors anything since he seems pretty comfortable thinking some pretty foul things without censoring them, but whatever.

now for the results! data breakdown under the cut

in total, of the swears I searched for, there's 286 swears across all four books. book two is the most vulgar, with 82 swears, and book three is the least, with only 52. book two also has the most variety (14 different words) and book three has the least variety (only 10 different words)

perhaps unsurprisingly, out of these 286 swears, SQQ's internal narration is responsible for 200 of them. his most used swear is "fuck" which appears 106 times in total and 86 times from SQQ (internal) specifically

following SQQ (internal) in the rankings is Shang Qinghua (internal) at 30 in total. his most used swear is "fucking." after that is SQH's external dialogue (14 swears, most used is "fuck") and then SQQ's external dialogue (11 swears, tied between "fuck" and "fucking"). considering how much he swears in his head, I think SQQ is exhibiting a surprising amount of restraint when it comes to swearing out loud!

aside from the transmigrators, the most vulgar side character is, surprisingly, Shen Jiu! he gets two internal uses of "fuck" and two external swears ("shitty" and "damn"). these are all from when he was younger, so perhaps this is a habit he broke when he got older and had to present a particular image as peak lord

the only other side characters who swear are Sha Hualing (internal), Bingge (internal), Linguang-Jun (external), Qiu Jianluo (external), and Ming Fan. all of them only swear once except Ming Fan, who gets two (one internal and one external)

of the 18 swear words I tracked, SQQ used all of them except one. the only person to say "bitch" is Ming Fan, out loud. #MingFanMisogynyMoments

SQQ does, however, have a monopoly on motherfuck/motherfucker/motherfucking, goddammit, dumbass, dick, and pussy. damn dude, you kiss your disciple with that mouth?

pre-transmigration, Shen Yuan swore three times in one forum post, those being "fucking," "damn," and "ass."

most usages of "ass" are for emphatic purposes, like "Relax, my ass!" Both SQQ and SQH really enjoy this construction lol

despite volume two being the most vulgar, it has the least variety of speakers! SQQ is responsible for all but two of those swears, the other two belonging to Sha Hualing and Ming Fan. volume four has the most variety, likely because it just involves a wide variety of characters

and that's my data roundup! go forth and use this as a resource when choosing how to insert swearing into your SVSSS dialogue <3

[science-bastard]

*lowers clipboard* ugh fine, you can kiss the other test subject for the 26th time. but just know that i’m dedicating an entire section of the research paper to this, and i’m gonna make it weird.

[useless19]

Let TFP!Megatronus be illiterate. Let him practise his speeches with Soundwave and have them make heavy use of Soundwave's ability to record/replay/stitch together sounds to get the best possible result. Let him have thousand-monkeys-on-a-thousand-typewriters-ed his way to revolutionary ideals.

So when Orion Pax shows up having read actual philosophy, it just drives a wedge between them, because it's all well and good to whitter on about pros and cons, but there's no decisive action suggested in the ramblings of old mechs. Theirs is about the argument, not the conclusion. And so, Orion comes across as condescending and Megatronus comes across as stubborn.

(And then sometime later, the closest thing to basic learning materials Megatron can get ahold of are medical/scientific notes and he discovers that he's got a knack for chemistry and immediately funnels that into plague development...)

[drfuckerm-d]

🥳🩷🛍️🎉

2 mini-playlists of the most fitting songs (opinion per moi) for two soong brothers, from my personal stash. the original playlists are 50+ songs but ive condensed them down to a more breathable 21 songs each. should u listen i hope u enjoy 🫶

[lorenzonuti]

Gathering data.

Twitter (X) | Instagram | Artstation

[nintendont2502]

the only valid ai is hal homestuck the rest of them can die by my blade

[humanoidhistory]

Computer center at Swedish State Railways office at Tomteboda, Stockholm, Sweden, July 21, 1965.

(digitaltmuseum)

[whyamiherewhosummonedme]

I have been informed that Reddit has found my story and are being super Reddit about it. Quick looksee of the comments revealed a bunch of people saying it was fake because [insert misunderstanding of how bacteriophages work]. Say its bullshit, whatever, but the being confidently absolutely wrong part is whats getting to me

[mostlysignssomeportents]

The reverse-centaur apocalypse is upon us

I'm coming to DEFCON! On Aug 9, I'm emceeing the EFF POKER TOURNAMENT (noon at the Horseshoe Poker Room), and appearing on the BRICKED AND ABANDONED panel (5PM, LVCC - L1 - HW1–11–01). On Aug 10, I'm giving a keynote called "DISENSHITTIFY OR DIE! How hackers can seize the means of computation and build a new, good internet that is hardened against our asshole bosses' insatiable horniness for enshittification" (noon, LVCC - L1 - HW1–11–01).

In thinking about the relationship between tech and labor, one of the most useful conceptual frameworks is "centaurs" vs "reverse-centaurs":

https://pluralistic.net/2022/04/17/revenge-of-the-chickenized-reverse-centaurs/

A centaur is someone whose work is supercharged by automation: you are a human head atop the tireless body of a machine that lets you get more done than you could ever do on your own.

A reverse-centaur is someone who is harnessed to the machine, reduced to a mere peripheral for a cruelly tireless robotic overlord that directs you to do the work that it can't, at a robotic pace, until your body and mind are smashed.

Bosses love being centaurs. While workplace monitoring is as old as Taylorism – the "scientific management" of the previous century that saw labcoated frauds dictating the fine movements of working people in a kabuki of "efficiency" – the lockdowns saw an explosion of bossware, the digital tools that let bosses monitor employees to a degree and at a scale that far outstrips the capacity of any unassisted human being.

Armed with bossware, your boss becomes a centaur, able to monitor you down to your keystrokes, the movements of your eyes, even the ambient sound around you. It was this technology that transformed "work from home" into "live at work." But bossware doesn't just let your boss spy on you – it lets your boss control you. \

It turns you into a reverse-centaur.

"Data At Work" is a research project from Cracked Labs that dives deep into the use of surveillance and control technology in a variety of workplaces – including workers' own cars and homes:

https://crackedlabs.org/en/data-work

It consists of a series of papers that take deep dives into different vendors' bossware products, exploring how they are advertised, how they are used, and (crucially) how they make workers feel. There are also sections on how these interact with EU labor laws (the project is underwritten by the Austrian Arbeiterkammer), with the occasional aside about how weak US labor laws are.

The latest report in the series comes from Wolfie Christl, digging into Microsoft's "Dynamics 365," a suite of mobile apps designed to exert control over "field workers" – repair technicians, security guards, cleaners, and home help for ill, elderly and disabled people:

https://crackedlabs.org/dl/CrackedLabs_Christl_MobileWork.pdf

It's…not good. Microsoft advises its customers to use its products to track workers' location every "60 to 300 seconds." Workers are given tasks broken down into subtasks, each with its own expected time to completion. Workers are expected to use the app every time they arrive at a site, begin or complete a task or subtask, or start or end a break.

For bosses, all of this turns into a dashboard that shows how each worker is performing from instant to instant, whether they are meeting time targets, and whether they are spending more time on a task than the client's billing rate will pay for. Each work order has a clock showing elapsed seconds since it was issued.

For workers, the system generates new schedules with new work orders all day long, refreshing your work schedule as frequently as twice per hour. Bosses can flag workers as available for jobs that fall outside their territories and/or working hours, and the system will assign workers to jobs that require them to work in their off hours and travel long distances to do so.

Each task and subtask has a target time based on "AI" predictions. These are classic examples of Goodhart's Law: "any metric eventually becomes a target." The average time that workers take becomes the maximum time that a worker is allowed to take. Some jobs are easy, and can be completed in less time than assigned. When this happens, the average time to do a job shrinks, and the time allotted for normal (or difficult) jobs contracts.

Bosses get stack-ranks of workers showing which workers closed the most tickets, worked the fastest, spent the least time idle between jobs, and, of course, whether the client gave them five stars. Workers know it, creating an impossible bind: to do the job well, in a friendly fashion, the worker has to take time to talk with the client, understand their needs, and do the job. Anything less will generate unfavorable reports from clients. But doing this will blow through time quotas, which produces bad reports from the bossware. Heads you lose, tails the boss wins.

Predictably, Microsoft has shoveled "AI" into every corner of this product. Bosses don't just get charts showing them which workers are "underperforming" – they also get summaries of all the narrative aspects of the workers' reports (e.g. "My client was in severe pain so I took extra time to make her comfortable before leaving"), filled with the usual hallucinations and other botshit.

No boss could exert this kind of fine-grained, soul-destroying control over any workforce, much less a workforce that is out in the field all day, without Microsoft's automation tools. Armed with Dynamics 365, a boss becomes a true centaur, capable of superhuman feats of labor abuse.

And when workers are subjected to Dynamics 365, they become true reverse-centaurs, driven by "digital whips" to work at a pace that outstrips the long-term capacity of their minds and bodies to bear it. The enthnographic parts of the report veer between chilling and heartbreaking.

Microsoft strenuously objects to this characterization, insisting that their tool (which they advise bosses to use to check on workers' location every 60-300 seconds) is not a "surveillance" tool, it's a "coordination" tool. They say that all the AI in the tool is "Responsible AI," which is doubtless a great comfort to workers.

In Microsoft's (mild) defense, they are not unique. Other reports in the series show how retail workers and hotel housekeepers are subjected to "despot on demand" services provided by Oracle:

https://crackedlabs.org/en/data-work/publications/retail-hospitality

Call centers, are even worse. After all, most of this stuff started with call centers:

https://crackedlabs.org/en/data-work/publications/callcenter

I've written about Arise, a predatory "work from home" company that targets Black women to pay the company to work for it (they also have to pay if they quit!). Of course, they can be fired at will:

https://pluralistic.net/2021/07/29/impunity-corrodes/#arise-ye-prisoners

There's also a report about Celonis, a giant German company no one has ever heard of, which gathers a truly nightmarish quantity of information about white-collar workers' activities, subjecting them to AI phrenology to judge their "emotional quality" as well as other metrics:

https://crackedlabs.org/en/data-work/publications/processmining-algomanage

As Celonis shows, this stuff is coming for all of us. I've dubbed this process "the shitty technology adoption curve": the terrible things we do to prisoners, asylum seekers and people in mental institutions today gets repackaged tomorrow for students, parolees, Uber drivers and blue-collar workers. Then it works its way up the privilege gradient, until we're all being turned into reverse-centaurs under the "digital whip" of a centaur boss:

https://pluralistic.net/2020/11/25/the-peoples-amazon/#clippys-revenge

In mediating between asshole bosses and the workers they destroy, these bossware technologies do more than automate: they also insulate. Thanks to bossware, your boss doesn't have to look you in the eye (or come within range of your fists) to check in on you every 60 seconds and tell you that you've taken 11 seconds too long on a task. I recently learned a useful term for this: an "accountability sink," as described by Dan Davies in his new book, The Unaccountability Machine, which is high on my (very long) list of books to read:

https://profilebooks.com/work/the-unaccountability-machine/

Support me this summer on the Clarion Write-A-Thon and help raise money for the Clarion Science Fiction and Fantasy Writers' Workshop!

If you'd like an essay-formatted version of this post to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:

https://pluralistic.net/2024/08/02/despotism-on-demand/#virtual-whips

Image: Cryteria (modified) https://commons.wikimedia.org/wiki/File:HAL9000.svg

CC BY 3.0 https://creativecommons.org/licenses/by/3.0/deed.en

[nostalgebraist]

On the topic of Declare:

A little while ago, on a whim, I prompted several different chat LLMs with a line from the Arabian Nights that has special importance in the novel:

O fish, are you constant to the old covenant?

In the book, this is used as a sign between members of a secret group, and there's a canned response that members are meant to give to identify themselves. (Like "I like your shoelaces" / "Thanks, I stole them from the President.").

I thought it'd be funny and impressive if one of these models responded with the canned phrase from Declare: that would have demonstrated both a command of somewhat obscure information and a humanlike ability to flexibly respond to an unusual input in the same spirit as that input.

None of the models did so, although I was still impressed with Gemini's reaction: it correctly sourced the quote to Arabian Nights in its first message, and was able to guess and/or remember that the quote was also used in Declare in follow-up chat (after a few wrong guesses and hints from me).

On the other hand, GPT-4 confidently stated that the quotation was from Jubilate Agno, a real but unrelated book:

When I asked Claude 3, it asserted that the line was from a real-but unrelated-poem called "The Fish," then proceeded -- without my asking -- to quote (i.e make up) a stanza from its imagined version of that poem:

It is always discomfiting to be reminded that -- no matter how much "safety" tuning these things are put through, and despite how preachy they can be about their own supposed aversion to "misinformation" or whatever -- they are nonetheless happy to confidently bullshit to the user like this.

I'm sure they have an internal sense of how sure or unsure they are of any given claim, but it seems they have (effectively) been trained not to let it influence their answers, and instead project maximum certainty almost all of the time.