#optimism bias
Explore tagged Tumblr posts
Text

https://twitter.com/JenniferLerner1/status/1698400322258104492
Just got to the part of Let This Radicalize You where they detail different responses to "major health threats" as outlined in The Psychology of Pandemics by Steven Taylor. Some ppl are "monitors", who seek out info "to cope with uncertainty."
"Meanwhile, others are prone to what's known as "blunting" behavior, which involves 'the distraction from, and minimizing of threatening information.'"
"Still others exhibit what's known as "unrealistic optimism bias," characterized by the belief that they are more likely than others to evade harm and experience positive outcomes."
[more on optimism bias]
"Research has shown that monitors are responsive to emotional appeals as well as detailed info abt risk factors & harm reduction strategies, whereas blunters are likely to avoid such messaging; for them, simple, logical messaging is likely most effective."
"Understanding these different reactions is essential to effective messaging and highlights the importance of taking multiple approaches in our organizing."
This thread is really putting into words that I think a major problem with communicating w/ loved ones who are "back to normal" wrt COVID is that many of us are using monitor messaging for folks who are more likely than not blunters.
#covid#psychology#optimism#optimism bias#bias#book#Let This Radicalize You#The Paychology of Pandemics#blunters#blunting#monitors
29 notes
·
View notes
Text
Understanding Trump Support: How Cognitive Biases Distort Reality for Many Voters
Reading time: 5 minutes Given the unmitigated disaster that the Trump's term was, it is baffling as to why he has polled so well and kept the race competitive. Three cognitive biases help us tell the tale, though.
SUMMARY: This post explores how cognitive biases—nostalgia, optimism, and proportionality—distort perceptions of Trump’s presidency and influence independent voter support. Nostalgia bias leads voters to remember Trump’s first term more favorably, while optimism bias makes them underestimate the potential hardships of another term. Proportionality bias causes them to dismiss concerns about…
#COVID19#Cognitive Bias#Distorted#Kamala Harris#Nostalgia Bias#Old FART#Optimism Bias#Proportionality Bias#Reality#Trump#Undecided Voters#Voters
1 note
·
View note
Text
still reeling from the ampers&one fanmeet so please enjoy this one moment that will keep me going until the end of eternity byeeee
#igm.talk#ampers&one#jiho#choi jiho#chey’s favs#idk why there’s no sound on desktop but there is on mobile??#also does anyone know how to optimize samsung s24 camera settings bc why was i 4 rows away from the stage#yet this looks like it was filmed on a potato#BUT ANYWAY BIG WIN FOR ME AS A JIHO BIAS OKAY BYE I’M INSANEEEEEEE ✨✨✨✨✨✨✨✨ DELUSIONNNN#PLS WATCH IT’S ONLY 9 SECONDS I’M SO UNWELLLLLLLL#JIHOOOOOOO 💖💖💖💖💖💖💖💖
13 notes
·
View notes
Text
The Illusion of Complexity: Binary Exploitation in Engagement-Driven Algorithms
Abstract:
This paper examines how modern engagement algorithms employed by major tech platforms (e.g., Google, Meta, TikTok, and formerly Twitter/X) exploit predictable human cognitive patterns through simplified binary interactions. The prevailing perception that these systems rely on sophisticated personalization models is challenged; instead, it is proposed that such algorithms rely on statistical generalizations, perceptual manipulation, and engineered emotional reactions to maintain continuous user engagement. The illusion of depth is a byproduct of probabilistic brute force, not advanced understanding.
1. Introduction
Contemporary discourse often attributes high levels of sophistication and intelligence to the recommendation and engagement algorithms employed by dominant tech companies. Users report instances of eerie accuracy or emotionally resonant suggestions, fueling the belief that these systems understand them deeply. However, closer inspection reveals a more efficient and cynical design principle: engagement maximization through binary funneling.
2. Binary Funneling and Predictive Exploitation
At the core of these algorithms lies a reductive model: categorize user reactions as either positive (approval, enjoyment, validation) or negative (disgust, anger, outrage). This binary schema simplifies personalization into a feedback loop in which any user response serves to reinforce algorithmic certainty. There is no need for genuine nuance or contextual understanding; rather, content is optimized to provoke any reaction that sustains user attention.
Once a user engages with content —whether through liking, commenting, pausing, or rage-watching— the system deploys a cluster of categorically similar material. This recurrence fosters two dominant psychological outcomes:
If the user enjoys the content, they may perceive the algorithm as insightful or “smart,” attributing agency or personalization where none exists.
If the user dislikes the content, they may continue engaging in a doomscroll or outrage spiral, reinforcing the same cycle through negative affect.
In both scenarios, engagement is preserved; thus, profit is ensured.
3. The Illusion of Uniqueness
A critical mechanism in this system is the exploitation of the human tendency to overestimate personal uniqueness. Drawing on techniques long employed by illusionists, scammers, and cold readers, platforms capitalize on common patterns of thought and behavior that are statistically widespread but perceived as rare by individuals.
Examples include:
Posing prompts or content cues that seem personalized but are statistically predictable (e.g., "think of a number between 1 and 50 with two odd digits” → most select 37).
Triggering cognitive biases such as the availability heuristic and frequency illusion, which make repeated or familiar concepts appear newly significant.
This creates a reinforcing illusion: the user feels “understood” because the system has merely guessed correctly within a narrow set of likely options. The emotional resonance of the result further conceals the crude probabilistic engine behind it.
4. Emotional Engagement as Systemic Currency
The underlying goal is not understanding, but reaction. These systems optimize for time-on-platform, not user well-being or cognitive autonomy. Anger, sadness, tribal validation, fear, and parasocial attachment are all equally useful inputs. Through this lens, the algorithm is less an intelligent system and more an industrialized Skinner box: an operant conditioning engine powered by data extraction.
By removing the need for interpretive complexity and relying instead on scalable, binary psychological manipulation, companies minimize operational costs while maximizing monetizable engagement.
5. Black-Box Mythology and Cognitive Deference
Compounding this problem is the opacity of these systems. The “black-box” nature of proprietary algorithms fosters a mythos of sophistication. Users, unaware of the relatively simple statistical methods in use, ascribe higher-order reasoning or consciousness to systems that function through brute-force pattern amplification.
This deference becomes part of the trap: once convinced the algorithm “knows them,” users are less likely to question its manipulations and more likely to conform to its outputs, completing the feedback circuit.
6. Conclusion
The supposed sophistication of engagement algorithms is a carefully sustained illusion. By funneling user behavior into binary categories and exploiting universally predictable psychological responses, platforms maintain the appearance of intelligent personalization while operating through reductive, low-cost mechanisms. Human cognition —biased toward pattern recognition and overestimation of self-uniqueness— completes the illusion without external effort. The result is a scalable system of emotional manipulation that masquerades as individualized insight.
In essence, the algorithm does not understand the user; it understands that the user wants to be understood, and it weaponizes that desire for profit.
#ragebait tactics#mass psychology#algorithmic manipulation#false agency#click economy#social media addiction#illusion of complexity#engagement bait#probabilistic targeting#feedback loops#psychological nudging#manipulation#user profiling#flawed perception#propaganda#social engineering#social science#outrage culture#engagement optimization#cognitive bias#predictive algorithms#black box ai#personalization illusion#pattern exploitation#ai#binary funnelling#dopamine hack#profiling#Skinner box#dichotomy
3 notes
·
View notes
Text
🫠🔪
#im so stupid#like i KNOW what happens#ive seen this 1000+ times#optimism bias is insane#god like wtf was i thinking#(i wasnt)#im going insane#apollos dodgeball hit me
4 notes
·
View notes
Text
finally started reading Poirot. I’m about 6 chapters into The Mysterious Affair at Styles. And I just want to point out that if Poirot is like retirement age (60sish?) and Hastings is 30, then when Hastings is having doubts about Poirot’s current methods while talking about how Poirot was a great detective “in his day”, “his day” would have been, what? Under the age of forty-five, at the very least? And Hastings speaks of it as if he was personally familiar with Poirot’s methods at the time.
This would mean Hastings was like thirteen or fourteen years old. Absolutely Tintining it up. And the fact that I am not yet aware of any adaptation where fortysomething Poirot is investigating crime scenes while tailed by his fourteen-year-old self-proclaimed apprentice from England is an immense tragedy.
#poirot#red randomness#hercule poirot#arthur hastings#also it makes hasting’s doubts in Styles way funnier#‘I don’t think he knows what he’s doing. he’s nowhere near as cool as he was when I was a Teen’#though of course Hastings has also just come back from the horrors of war#and has probably lost much of his optimism and faith in everything#even as he keeps his tone in writing reasonably upbeat#but yes. small boy hastings determined that Poirot needs to see him as a Very Grown Man.#imagine.#(I don’t see it in a father figure way though. partially because I have a Certain Bias)#(but also just because weird intergenerational friendships are always extremely fun to me)#(my boyfriend said it reminded him of Ace Attorney and he’s right)
6 notes
·
View notes
Text
Advanced Methodologies for Algorithmic Bias Detection and Correction
I continue today the description of Algorithmic Bias detection. Photo by Google DeepMind on Pexels.com The pursuit of fairness in algorithmic systems necessitates a deep dive into the mathematical and statistical intricacies of bias. This post will provide just a small glimpse of some of the techniques everyone can use, drawing on concepts from statistical inference, optimization theory, and…

View On WordPress
#AI#Algorithm#algorithm design#algorithmic bias#Artificial Intelligence#Bayesian Calibration#bias#chatgpt#Claude#Copilot#Explainable AI#Gemini#Machine Learning#math#Matrix Calibration#ML#Monte Carlo Simulation#optimization theory#Probability Calibration#Raffaello Palandri#Reliability Assessment#Sobol sensitivity analysis#Statistical Hypothesis#statistical inference#Statistics#Stochastic Controls#stochastic processes#Threshold Adjustment#Wasserstein Distance#XAI
0 notes
Text
From Data Dependence to Informed Insight
In the complex realm of supply chain management, understanding biases can be crucial for effective decision-making. In this episode of the Chain Reaction Podcast, we explore the phenomenon of quantitative bias, particularly as highlighted in Linda Chang et.al., insightful paper, “Does Counting Change What Counts?” Quantitative bias encompasses systematic errors in research, particularly in…

View On WordPress
#Confounding Bias#Data Dependence#Decision Making#Qualitative Research#Quantitative Bias#Research Bias#Research Biases#Supply Chain Management#Supply Chain Optimization
0 notes
Text
Optimism Bias - Idea to Value
Just how likely are you to succeed? And how unlikely is it that something bad happens to you, just because it happened to someone else? Surprisingly, most people appear to think that in the future, it is more likely that good things will happen to them, and less likely that bad things would happen to them. This is a cognitive bias known as the optimism bias. According to researchers, this is one…

View On WordPress
0 notes
Text
In theory, I am wildly underpaid
In theory, I am wildly underpaid I already know that the Comprehensive Financial Planning service can literally transform the way that a household looks at its financial future. I have receipts, and I am pretty sure clients will say so, publicly. At $3000 for a lifetime of financial guidance, it is wildly underpriced. This underpricing will end on December 31, 2023. In practice… People think…

View On WordPress
#healthcare#Lifetime Income#investment strategy#portfolio#Roth IRA#AI#Bias#Investment#Insurance#Comprehensive Financial Planning#Guidance#Theory#Financial#Underpricing#Social Security#Free Tools#Financial Planning#Tax Optimization
0 notes
Text
Sure, but I still want to know their priors and sampling techniques. Failing that, using the most easily accessible methods to gather data can still yield potentially interesting information about overall dynamics even if we apply mathematical analyses that assume oversampling of queer users to "correct" the effects of snowball sampling. It's worth noting that sampling information about human sexuality is pretty much uniformly nightmarish in any case; this is actually not that much worse than published peer reviewed sampling efforts, horribly enough.
I am taking everyone who made a poll to gauge the True Percentage of Queers on Tumblr and putting them through a statistics course
#if i had a nickel for every psych study focusing on queer people that uses snowball sampling for their queer sample#and randomly harvested psych 101 undergrads for their “heterosexual” sample#pretty sure there's a buck in there just from my asexuality studies folder alone#admittedly we're not trying to compare two samples here we're trying to estimate prevalence#but real scientists are also likely to use nonoptimal sampling techniques in experiments if optimal ones are deemed Too Hard#my favorite is “haphazard sampling” == “we have some population of little critters and we grabbed the first n we could catch as our sample”#(it's not random because catchability is a metric of the individual little dudes in question)#anyway in conclusion if they're not going to put their methods section in plain English I'm treating them like a Science paper#and assuming shit is significantly more dodgy than it looks at first blush#at least I know what the sampling bias is on this version and if it gets big enough I can use specific tools to deal with it#even if that is simply comparing number of active accounts to number of responding accounts and assuming various levels of bias response#if we have clear methods we can do math about the obvious flaws#if we don't all we can do is imagine the way we hope data was collected#and look I've been in this game long enough to be Skeptical#also that said: this is largely for fun and sarcasm
44K notes
·
View notes
Text
Mamamoo's Fuck-Fitness Program

(Male Reader x Hwasa & Solar, 5k Words) Tags: Workout sex; Sex while working out; Someone grows a GirlCock (Blame Dreamcatcher); Vagina, Oral, Anal, and Exterior Sex; Double penetration; Multiple cocks in one hole; Sweaty messy sex; Only Hwasa and Solar are there though, sorry; Sloppy Seconds; Lots of calories were burned during this smut; Can recommend this exercise for weight-loss; Thigh-fucking
Sex, is a messy activity. The grunting and exhalations, the sordid stench of sweat and hormones, the noisy slap of flesh, the taste of another person upon your tongue, the sticky fluids that invariably end up leaking everywhere... Sex was an exercise drenched in shared shamelessness, the thrill of ignoring your innate disgusts to focus solely upon the pursuit of pleasure; and of course, breeding. Which only made threesomes all the nastier. Now there was a third body squishing itself against the others, adding its own sensual heat to the mix, spicing the intercourse with its uneven dynamic that only ever encouraged more perversion. It was simply too easy to toss away your inhibitions and give in to the unnatural deviancy of the situation, spurring yourself as you seek your climax. Which, all in all, meant that you burned a lot more calories in a threesome, which was why the Mamamoo Workout Program always made sure its clients were paired with two of their ladies at once!
Fans would obviously plumb the depths of their reserves when with one of their idols, but with two of them... why, the weight would practically melt away!
It cost quite a bit to secure a session at the Mamamoo Fitness Center, but it was well known that the health benefits were worth every penny. There were only four open slots per day, and competition was fierce to obtain one, it was said that the ladies were fully booked for the year within minutes of opening; it was something of a miracle you had managed to secure yourself one. The modest office itself was tastefully decorated, pure white walls covered in informational posters and awards, and the plush chairs in the waiting room were of the highest quality. The assistants behind the front desk were equally as beautiful as the room itself, and went about their business with the quiet diligence of any medical office. A smiling secretary had given you a clipboard of forms and waivers to fill out, full of disclaimers carefully worded to avoid any explicit mention of what went on. The form had blandly informed you that your upcoming session would consist of a "personalized workout session", guided by two randomized members of Mamamoo, who would lead your "workout" until you had reached your "fitness goals". What it really meant of course, was that you would be vigorously fucked until you were on the edge of exhaustion.
The Mamamoo Workout experience had been carefully tuned to maximize the amount of calories burned during the intercourse. The controversial decision to randomize the members chosen had been done on purpose, since it had been noted that clients often burned as many calories with their least favorite member as they would have with their bias. One particularly celebrated patient had spent five hours rutting atop Wheein, and had managed to lose 20 pounds in the process. The eight hour long slots had also been chosen as the most optimal amount of time, past that and most patients were on the verge of death via excessive fluids loss. And of course, the decision to include two of the members instead of just one was to send their clients into a sexual frenzy; though it was also because the members enjoyed having sex with one another too much. On average, most patients lost around ten pounds over the several hours of intensive intercourse, their excess fat burning away as they pushed themselves to the limit to continue fucking half of Mamamoo. It was an extremely efficient method of losing weight, with the added benefit of being the best sex of your entire life.
The demure secretary returns to collect you, and your documents, checking to make sure everything had been filled out properly before guiding you to the front desk to record your temperature and weight. Nodding with satisfaction, she leads you out of the waiting room and through the doorway in the back, into a plain corridor adorned only with four brown doors. Colorful placards on the doors indicate whether the rooms were in use or not, and as you pass by the single occupied suite, you hear muffled squeals and moans emanating from it; evidently the sound-proofing here was high-quality. The assistant takes you to the room two doors down from the one currently in use, politely knocking on the door before opening it and motioning you inside. Your heart pounds and your palms grow sweaty as the reality of your situation hits you, you had been anticipating this for so long it had become almost mundane, but now that you were on the threshold... The secretary gives you an encouraging slap on the ass, and smiling kindly, assures you that you would be up for the grueling workout ahead of you; everyone gets cold feet before this! Mustering your courage, you return the favor, much to her amusement, before entering the "fitness room"; the door shuts, and locks, behind you.
Twin goddesses await you within, hands on their hips as they appraise their newest customer, their bodies sheathed in clinging tights that accentuated their plush lower lips, and sports bras that pushed out their chests. Solar seemed as bright as her namesake, cheerfully eyeing your crotch with open intent, while Hwasa glares thoughtfully at your face, her own sultry expression hinting of her love for vigorous copulation. Solar steps forward to greet you, her lithe body swaying as takes your hand and leads you further into the exercise room; which was severely lacking in conventional exercise equipment. The floor was completely covered in firm mats, with beanbags and exercise balls strewn about, there were several curiously-shaped benches, and the pull-up stand had a few too many straps than usual. Full-length mirrors covered the two sidewalls, while the furthest had various cabinets and a door that presumably led to the bathroom; and there was a tallyboard that marked the gender of every customer, it seems that women were here as often as men. Hwasa saunters up and takes your other hand, smirking slightly as she senses your nervousness, and you stammer as you introduce yourself to the two idols. The pair smile at your awkwardness, before explaining today's schedule to you once more; a mild stretching session, followed by a series of intensive workouts interspersed with breaks for hydration and restroom use, with a shower at the end to clean you all off.
Contrary to your assumptions, your time with Hwasa and Solar began with the utmost banality, the pair guiding you through some basic standing stretches before moving onto some sitting ones. If the police had burst into the room under suspicion of illegal sexual acts, they would have found a normal guy in sweats and a baggy tshirt, grunting as he strained to touch his toes, while two attractive fitness instructors watched with amusement. Now if said police returned half an hour later, well... The first sign that this was anything other than a regular weight-loss session was when Hwasa moved in front of you while your legs were spread, and smoothly slid her ass back between them. A position that left you quite conscious about the size of her posterior, now that it was nuzzling against your crotch, the pair's chests might have been similar, but Hwasa's rear was far thicker. A moment later Solar presses herself against your back, squishing her breasts against you as she slowly leans forward, stretching you out while ensuring you would be unable to escape. If Hwasa is at all troubled by the now obvious poking her cheeks, she does not show it, and placidly starts to grind against your loins; your penis obviously required stretching out as well. What started as slow teasing gradually turns into something more, as your excitement builds you find yourself moving as well, until what started as a suspicious stretch has degenerated into desperate dry-humping.
You groan as you rub yourself against Hwasa's thick ass, boldly grasping her waist so as to force her back against you all the harder, before long you are leaking through your sweatpants and onto her tights. She slowly hikes her ass up, and you eagerly follow, until it looks as if you were plowing her doggy-style, only with more clothes than usual on. Speaking of which, now that you have the room, you eagerly pull down your pants, and are surprised to discover your shirt getting yanked up as well, before Solar meshes herself against your back once more. Her hands roam your bare chest, toying with your nipples while she kisses and sucks your neck, whispering encouragement into your ear before nibbling up on it. You shudder as you hotdog Hwasa's plump butt, the fabric of her tights smooth enough to allow for easy movement, the sticky evidence of your virility smearing itself against it. With her amused face staring back at you in the mirror, the idol suddenly pulls forward, using her hand to press your cock downward before pushing back against you; and you find your cock sandwiched between her meaty thighs. Now it truly looked as if you were having sex, as you furiously fuck Hwasa's thighs, your crotch slapping loudly against her constrained ass as she squishes her legs together in imitation of her tight pussy.
With Solar pressing her supple body against you from behind, and with your cock buried between Hwasa's thighs, you feel your climax building unstoppably. The pair of idols were not blind to this, and they eagerly brought about its fruition. Solar smiles wickedly at you in the mirror, "Finishing so soon, dear?" she purrs, "we haven't even completed our stretches yet," before returning to giving your neck purple hickeys. Hwasa simply looks back at you and commands, "Cum, now," and you moaningly obey, picking up the pace as she clenches around your manhood. You had not even lasted long enough to put your first load inside one of the members of Mamamoo, and it is with some embarrassment that you empty yourself against Hwasa's stomach. You groan as you paint her tummy with your seed, your cock fooled into thinking that it was draining itself inside of her fertile pussy, when instead your sperm was wasted upon her tanned skin. Shaking from the force of your orgasm, you lean on Hwasa for support as your strength pours from your dick, as the idols loudly praise your climax.
Solar pulls you upright onto your feet, murmuring encouragingly as she massages your crotch and thighs, while Hwasa idly cleans your semen off of her stomach with her fingers, licking each in turn while staring bemusedly at you. The girls share a glance before Hwasa gets up and stalks over to a cabinet, while Solar occupies you with kisses and touches, now that the warmup was finished, it was time to pick up the pace. And of course, the pair knew the best way to encourage their clients to do just that. Mamamoo, like most Third-Generation idols, may have looked on with contempt at the alchemical aphrodisiac brews churned out by Twice, but that did not mean they were opposed to all such potions. Hwasa removes an ornate, maroon-colored flask from a drawer, tossing it to Solar before sauntering back and shamelessly stripping in front of you. Your eyes boggle as you drink in the sight of Hwasa's naked body, her fertile curves framing a massively puffy pussy that is already slick with arousal; and her swaying breasts were practically begging to be groped. Meanwhile, Solar takes advantage of your distraction and chugs a portion of the bottle, licking her lips as the cherry-flavored drink pours into her stomach and its effects take hold.
Solar lets out a moan of pure pleasure as a bulge appears in her tights, spreading upwards until it nearly pushes fully from beneath the fabric. She shudders as she gingerly reveals the results of the potion, a girl-cock of average size, twitching and tender from its recent birth, with a pair of heavy balls throbbing in their hairless sack. Mamamoo know very well that to inspire a client, it is best to set the pace yourself, and let men's natural jealousies do the rest; plus, the women tend to be rather invigorated by having their favorite idol's cock in one or more of their holes. So Hwasa wastes no time in getting on her knees in front of Solar, and enthusiastically starting to suck her off. You are admittedly shocked by this sudden development, you had not expected for one of Mamamoo to sprout a cock, but you are not entirely opposed to this... To your own surprise, you find yourself oddly aroused by the scene taking place in front of you, and soon are shoving your own stiffening cock into Hwasa's face. The idol expertly gives oral to both of you, one hand on either cock, sucking one then the other, rubbing the tips together as her tongue lathers both with warm spit. You and Solar both moan from Hwasa's efforts, kissing and groping one another as your dicks are devoured, precum dribbling steadily all over her pert lips. After what seems like an age of endless pleasure, Hwasa smoothly rises before bending over, her face still bobbing in front of your crotches, but now offering an alternative.
You are too busy enjoying the stimulation of getting a blowjob from an idol like Hwasa, and so Solar is the first to take her pussy, slipping inside of her groupmate with practiced ease. While Solar has her sloppy way with Hwasa's lower lips, you use her upper lips with equal messiness, using her head's positioning to force your cock down her throat. Hwasa gags as you fuck her face like a pussy, slobber coating your balls with every thrust as you struggle to not unload your seed directly into her stomach. You and Solar hold hands and make-out while you spit-roast Hwasa, who squirms as both of you shove yourselves as deep as possible into her, united in your eagerness to use Hwasa's body. But before you can impregnate Hwasa's guts, Solar does just that inside of her friend's cunt, moaning loudly and spasming wildly as she empties herself into Hwasa, her eyes rolling back as pleasure washes through her body. Panting, her tongue still connected to yours be streamers of spit, Solar pulls out of Hwasa, slapping her thick ass in thanks before wiggling her eyebrows suggestively at you; now its your turn... Hwasa solidly plants herself against one of the wall mirrors in preparation for the pounding she was about to receive, looking back smugly at you as she spreads one cheek open in welcome. You shudder as you press your cockhead into the sticky mess leaking from Hwasa's plump pussy, your passage into her dripping hole made even easier by Solar's seed.
The cacophony of flesh slapping together reverberates through the exercise room as you violently take Hwasa against the wall. A surprising vigor fills you, turning every thrust into a hammer-blow as you break yourself against her massive ass-cheeks, your cock churning her pussy as Solar's seed pours out of it. Her sloppy cunt constricts tightly around your shaft with every thrust, gripping you tightly even while taking a ferocious pounding; idol pussy truly was superior. Hwasa growls hungrily as you plow her, licking the mirror lustfully to spur you on, her cunt gushing as she climaxes from the intensity of your coupling. With your hands around her hips, you hold her steady as you go into a frenzy, thrusting wildly as your own orgasm approaches, gasping her name with every breath while your balls throb with effort. Sweat is pouring down your skin when you finally let out a groan and force yourself deep inside of Hwasa, your semen erupting into her in a flood of jizz. Who looks back at you with satisfaction, licking her luscious lips while you plaster her pussy with your seed; relishing in every drop coating her insides. When you are finished, you stagger backwards, exhausted by your sex, and a laughing Solar passes you a water bottle as she guides you to the bathroom for a little break.
You were perplexed by your energetic coupling with Hwasa, you had not intended to be so rough with her, even though she had evidently enjoyed it, and you ponder this while you piss in one of the two toilets with Solar. Hwasa soon joins you two, squatting on the free seat and voiding her bladder as well; the pair of idol's did not need to ruminate on your sudden vigor, they knew exactly where it came from. Like most idols, Mamamoo knew full well that when fucking a cum-filled hole, men naturally grew more forceful and energetic, their bodies automatically adapting its rhythm to better scoop the semen out with its thrusts; which of course, burned more calories. So, the girls made sure to have as many holes filled with their own girlcocks and cum as soon as possible, not that it was very hard considering how good those holes were; and jealousy was a truly powerful motivator. Thus, when you three exit the bathroom, Hwasa almost immediately is laying down on padded bench, and opening her legs for Solar, who happily starts fucking your sloppy seconds out of Hwasa. You meanwhile rest on an exercise ball, idly rubbing your still slick yet flaccid cock, while the two idols make passionate love without you; a situation that makes you surprisingly aroused as you watch them fucking. Eventually, the squeals and moans are enough to get you hard once more, and you hasten to join the pair of idols once more, already knowing which hole you wanted to fill.
Solar starts with surprise when your hands grasp her waist, stopping mid-thrust to glance back at your touch, bemusement written on her features, "Oh, would you like a turn?" she asks with a knowing smile, before her eyebrows raise as you apply pressure to guide her down onto Hwasa, "Oh, wait, are you lubed up enough...?" Solar's question is answered as you press your glistening cock against her exposed asshole, and force yourself inside of her. Solar groans as she is impaled from behind, pushing herself deeply into Hwasa as you gradually fill her ass; until you are both balls-deep in your respective holes. Solar squirms as she is pleasured from both sides, her girl-cock getting massaged by Hwasa's sultry cunt, while your dick pokes at her innards; and you had not even started thrusting yet. Solar's guts probably had subtle differences compared to Hwasa's experienced pussy, but by now you were on your third load, and all your cock cared about was that it was inside of a warm hole; so it does what any dick inside of a warm hole would do, and starts moving. After an awkward start, all three of you eventually fall into a steady rhythm, allowing both you and Solar to get deep strokes in, while Hwasa simply lays there and takes it, naughtily urging you both on while you do all the work. Solar's perky butt meshes perfectly against your crotch with every thrust, and you have to contain yourself to not simply pin her against her groupmate and pound her soggy guts out of shape. Solar though, sounds like she is in heaven, gasping and moaning, becoming overstimulated and giving into her own pleasure, frantically humping away between Hwasa's thick thighs until she reaches her climax. When she pushes deep inside of her fellow idol, you follow suit, pressing yourself against her sweaty back as your weight presses Solar's girl-cock even further into Hwasa's pussy. Shuddering, Solar collapses onto her and unloads the contents of her balls into Hwasa, who groans as the warmth spreads through her belly, squished beneath both of your weight.
A glistening streamer of semen connects the tip of Solar's cock with the mess leaking out of Hwasa's cunt as she delicately maneuvers out from between you two; leaving you yet another sticky mess to unclog. But Hwasa has other ideas; obviously tired of being passive, instead she cranks the back of the bench upwards and seats you on it, sperm and sweat streaking down her thighs as she gazes down at you. She squats down in front of your upraised dick and takes it in her mouth, slobbering on your smelly meat until it is thoroughly doused in spittle, her eyebrows furrowed intensely as your tip stabs at her throat. Hwasa wears a seductive smirk when she finishes lubricating your manhood, rising back up to straddle you, placing one hand on your chest to hold you down, she uses the other to guide your cock inside of her as she sits on it. Your eyes widen in surprise when instead of slipping easily into her slimy cunt, Hwasa angles you a little further back, and forces your dick inside of her unused asshole. She lets out a satisfied grunt as she hilts you, your member twitching inside of her belly, her lips quirking upwards when she notices your expression, "What, did you think I disliked having my ass filled as well?", she brings a finger to your lips to hush you, "No, just shut up and let me enjoy myself." So you do, and she does, though it was not as if you were not savoring the feeling of Hwasa's tight anus sliding up and down your shaft as much as she was.
Hwasa's powerful thighs piston her up and down your length, a sneer of pleasure twisting her lips as your cock grinds against her insides, making you shudder with every squat. If anything, she looked smug at being able to exert some control once more, leaning down to forcefully kiss you on whim, her cunt leaking her juices steadily onto your belly. Moaning, you grope her swaying breasts as Hwasa rides you, holding onto them for support as her asshole clamps tightly around your cock. Only when you approach the edge of your orgasm does she pause, waiting until your dick finishes pulsing before resuming her sensual movements, her hips writhing atop your lap as she works you deep inside of her. After Hwasa has successfully edged you for a third time, Solar reappears behind her, no doubt painfully erect once more, running her hands down her friend's back to attract her attention. But Hwasa is teasingly aloof, "Sorry dear, this hole is already taken," she purrs, grinding meaningfully upon you for emphasis, curling her mouth into a haughty yet knowing smile. Solar simply beams though, and scoots herself onto the small open portion of the seat slipping her legs up yours until her cock and balls are squished against your own. Rolling her eyes, Hwasa stands up enough until only your head is inside of her, and you twist to try and see what is going on as you feel something hard pressing against the base of your tip. The pressure builds, and only when another rigid object surges in alongside your cock do you realize that Solar has forced her dick into the idol's ass as well. Hwasa lets out a true groan as she sits on both of your cocks, clenching tightly around both of your shafts as precum leaks out from the gaps between your cocks. With frightening flexibility, she lifts one leg up over you and turns, so that her side is now facing you, and you can see Solar smiling reassuringly from across Hwasa's swollen breasts; Hwasa's ass felt so snug with someone else in it...
With a hand on either shoulder, Hwasa bounces lustfully on both of your cocks, leaving both you and Solar moaning plaintively. Effectively pinned by each other's weigh, Hwasa has free reign to fuck you as she liked, turning what should have been in intense anal pounding into a languidly intimate ride. With your shafts squishing and slipping against one another, Hwasa's guts massage your dicks until they are the edge of bursting, then she pauses, allowing your precum to slop out over your congealed cockheads, before resuming her exquisite torture. All you and Solar can do is grab at her curvaceous body as it pumps up and and down your members, and your mind starts to melt under the unending pleasure. Hwasa's asshole makes disgusting squelching noises as precum slops out of it and onto your balls, soon your lower bodies are splattered with the evidence of your weakness. Even an experienced slut like Solar looks to be in heaven, her tongue lolling as she glares lustfully back at you, grinding herself against you in her desperation to climax. You start asking for it, pleading with Hwasa for release, which of course only makes her leak all the more, until eventually she lets out a grunt a paints the matting with a wash of gooey squirt, shuddering with ecstasy as her asshole clamps tightly around your members. Her face flushed with arousal, her hands slip onto your throats as she snarls, "Beg for it! Beg. For. It!" and with your brains sloshing with hormones, you and Solar easily comply. You beg for release as Hwasa slams herself against you, her thick ass cheeks slapping loudly as fluids spray, the two of you now humping wildly upwards while Hwasa hammers you flat again and again. Solar is the first to finish, moaning sweetly, her eyes roll back as she fills Hwasa's guts, her pulsating cock slobbering cum all over your dick as much as her coils. The stimulation is enough to make you explode soon after, drowning Hwasa's innards in semen as she squirts messily once more as her ass gets turned into a jizz-filled slurry.
Hwasa unmounts you both with a groan, Solar nearly toppling backwards off the seat as all three of you are weak and trembling from your intensive exertions. Hwasa gingerly kneels next to you, and without hesitation starts cleaning off the filthy mess coating your sensitive cocks with her mouth. Now you are truly writhing in the seat, your skin painfully stuck to the black leather by sweat, clutching her hair as her warm saliva is slathered all over your manhood. Hwasa skillfully rubs your tips together, her tongue slipping between and around them, before taking you both in her mouth, trying to fit both of your cocks into her throat and sometimes succeeding. By the end of it, both you and Solar are gasping, but your cocks are both bulging once more. Hwasa stands back up before sashaying to an uncluttered area of mat and bending over, spreading her cheeks for your enjoyment. She smirks coyly as her asshole belches a ream of semen down her leg, her gaping pussy still awash with sperms, "Who is getting which hole?", she asks, her smile growing wider, "Or are you going to share again...?" Hwasa seductively licks her lips as Solar untangles herself from you and staggers over, with you a step behind, wiggling her butt enticingly as you approach. Solar glances at you and gives you a kiss, before grinning, "Well? We have five hours left, so take your pick, or do you want me...?"
The next four hours pass in a blur of pounding flesh and spewing fluids, an unending tide of hedonism that sees you plumbing the depths of your depravity. You violate Hwasa's asshole, you plow her pussy, you fill her throat, while Solar gleefully does the same. You balls slap together in an endless cacophony, and more often than not your shafts slip and squelch against each other as you both ravage the same hole at once. Not that Solar's orifices are spared either, pumping away at her butt while Hwasa pleasures you both, or forces you to pleasure her; and Solar's face is almost as covered with both her cum and your own as Hwasa's is. All of you guzzle down vast amounts of water, only stopping when your bodies can take it no more and you have to make a break to the bathroom, often returning to find the other two fucking as intensely as before. Sweat pours down your skin like a waterfall, and the mats are covered with puddles of sexual fluids (and the occasional pond of piss from when someone was unable to reach the toilet in time), it feels like your brain has devolved until your only thoughts are for slaking your desires upon the two idols of Mamamoo. Eventually though, your exhaustion catches up to you, and when you pull out of Hwasa's sloppy cunt for what seems like the hundredth time that day you find yourself collapsing from utter fatigue.
Hwasa and Solar carefully tend to you over the next twenty minutes, pouring fluids down your throat to rehydrate you and massaging your cramped and aching muscles. When they are satisfied you can walk, they haul you to your feet, chattering amiably and complimenting you on your love-making skills; of course they barely showed even a hint of tiredness. They haul you into the bathroom, where they crudely hose the worst of the filth off of you with a shower head before frog-marching you to the exit. A smiling assistant greets you at the door, and your half of Mamamoo wave you goodbye as she leads you out, covered head to toe in cum, squirt, and other unmentionable fluids. As the girl leads you down the corridor, you idly notice that the idols had stuffed you back into your clothes, though you knew you must absolutely reek. The other occupied door cracks open as you approach, and a lady staggers out of it, her belly distended and her holes leaking so much cum it is slicking through her yoga pants in a reeking stream. She barely even glances at you as you are dragged by her, though through the closing door you notice a rather well-endowed Moonbyul and Wheein blowing kisses to their retreating client. Humming cheerfully, the assistant takes your weight at the scale, before depositing you in front of the kindly secretary from before. Who beams knowingly, before slyly enquiring if you would like to make an appointment for next year? Nodding tiredly, you flick through your calendar to look up the date, while the lady behind you noisily vomits what sounds like gallons of semen into a trashcan.
Of course you would like to make another appointment, though you are sure to ask if there have been any last minute cancellations... Another few more sessions like this, and you would be hitting your weight-loss target in no time!
497 notes
·
View notes
Text
Imma just jack off about it bc theres not much i can do. I might window shop around for hands free toys bc my wrists hurt. I aint got no money, literally 66 cents to my name but i can dream 😔
0 notes
Text
Direct Preference Optimization: A Complete Guide
New Post has been published on https://thedigitalinsider.com/direct-preference-optimization-a-complete-guide/
Direct Preference Optimization: A Complete Guide
import torch import torch.nn.functional as F class DPOTrainer: def __init__(self, model, ref_model, beta=0.1, lr=1e-5): self.model = model self.ref_model = ref_model self.beta = beta self.optimizer = torch.optim.AdamW(self.model.parameters(), lr=lr) def compute_loss(self, pi_logps, ref_logps, yw_idxs, yl_idxs): """ pi_logps: policy logprobs, shape (B,) ref_logps: reference model logprobs, shape (B,) yw_idxs: preferred completion indices in [0, B-1], shape (T,) yl_idxs: dispreferred completion indices in [0, B-1], shape (T,) beta: temperature controlling strength of KL penalty Each pair of (yw_idxs[i], yl_idxs[i]) represents the indices of a single preference pair. """ # Extract log probabilities for the preferred and dispreferred completions pi_yw_logps, pi_yl_logps = pi_logps[yw_idxs], pi_logps[yl_idxs] ref_yw_logps, ref_yl_logps = ref_logps[yw_idxs], ref_logps[yl_idxs] # Calculate log-ratios pi_logratios = pi_yw_logps - pi_yl_logps ref_logratios = ref_yw_logps - ref_yl_logps # Compute DPO loss losses = -F.logsigmoid(self.beta * (pi_logratios - ref_logratios)) rewards = self.beta * (pi_logps - ref_logps).detach() return losses.mean(), rewards def train_step(self, batch): x, yw_idxs, yl_idxs = batch self.optimizer.zero_grad() # Compute log probabilities for the model and the reference model pi_logps = self.model(x).log_softmax(-1) ref_logps = self.ref_model(x).log_softmax(-1) # Compute the loss loss, _ = self.compute_loss(pi_logps, ref_logps, yw_idxs, yl_idxs) loss.backward() self.optimizer.step() return loss.item() # Usage model = YourLanguageModel() # Initialize your model ref_model = YourLanguageModel() # Load pre-trained reference model trainer = DPOTrainer(model, ref_model) for batch in dataloader: loss = trainer.train_step(batch) print(f"Loss: loss")
Challenges and Future Directions
While DPO offers significant advantages over traditional RLHF approaches, there are still challenges and areas for further research:
a) Scalability to Larger Models:
As language models continue to grow in size, efficiently applying DPO to models with hundreds of billions of parameters remains an open challenge. Researchers are exploring techniques like:
Efficient fine-tuning methods (e.g., LoRA, prefix tuning)
Distributed training optimizations
Gradient checkpointing and mixed-precision training
Example of using LoRA with DPO:
from peft import LoraConfig, get_peft_model class DPOTrainerWithLoRA(DPOTrainer): def __init__(self, model, ref_model, beta=0.1, lr=1e-5, lora_rank=8): lora_config = LoraConfig( r=lora_rank, lora_alpha=32, target_modules=["q_proj", "v_proj"], lora_dropout=0.05, bias="none", task_type="CAUSAL_LM" ) self.model = get_peft_model(model, lora_config) self.ref_model = ref_model self.beta = beta self.optimizer = torch.optim.AdamW(self.model.parameters(), lr=lr) # Usage base_model = YourLargeLanguageModel() dpo_trainer = DPOTrainerWithLoRA(base_model, ref_model)
b) Multi-Task and Few-Shot Adaptation:
Developing DPO techniques that can efficiently adapt to new tasks or domains with limited preference data is an active area of research. Approaches being explored include:
Meta-learning frameworks for rapid adaptation
Prompt-based fine-tuning for DPO
Transfer learning from general preference models to specific domains
c) Handling Ambiguous or Conflicting Preferences:
Real-world preference data often contains ambiguities or conflicts. Improving DPO’s robustness to such data is crucial. Potential solutions include:
Probabilistic preference modeling
Active learning to resolve ambiguities
Multi-agent preference aggregation
Example of probabilistic preference modeling:
class ProbabilisticDPOTrainer(DPOTrainer): def compute_loss(self, pi_logps, ref_logps, yw_idxs, yl_idxs, preference_prob): # Compute log ratios pi_yw_logps, pi_yl_logps = pi_logps[yw_idxs], pi_logps[yl_idxs] ref_yw_logps, ref_yl_logps = ref_logps[yw_idxs], ref_logps[yl_idxs] log_ratio_diff = pi_yw_logps.sum(-1) - pi_yl_logps.sum(-1) loss = -(preference_prob * F.logsigmoid(self.beta * log_ratio_diff) + (1 - preference_prob) * F.logsigmoid(-self.beta * log_ratio_diff)) return loss.mean() # Usage trainer = ProbabilisticDPOTrainer(model, ref_model) loss = trainer.compute_loss(pi_logps, ref_logps, yw_idxs, yl_idxs, preference_prob=0.8) # 80% confidence in preference
d) Combining DPO with Other Alignment Techniques:
Integrating DPO with other alignment approaches could lead to more robust and capable systems:
Constitutional AI principles for explicit constraint satisfaction
Debate and recursive reward modeling for complex preference elicitation
Inverse reinforcement learning for inferring underlying reward functions
Example of combining DPO with constitutional AI:
class ConstitutionalDPOTrainer(DPOTrainer): def __init__(self, model, ref_model, beta=0.1, lr=1e-5, constraints=None): super().__init__(model, ref_model, beta, lr) self.constraints = constraints or [] def compute_loss(self, pi_logps, ref_logps, yw_idxs, yl_idxs): base_loss = super().compute_loss(pi_logps, ref_logps, yw_idxs, yl_idxs) constraint_loss = 0 for constraint in self.constraints: constraint_loss += constraint(self.model, pi_logps, ref_logps, yw_idxs, yl_idxs) return base_loss + constraint_loss # Usage def safety_constraint(model, pi_logps, ref_logps, yw_idxs, yl_idxs): # Implement safety checking logic unsafe_score = compute_unsafe_score(model, pi_logps, ref_logps) return torch.relu(unsafe_score - 0.5) # Penalize if unsafe score > 0.5 constraints = [safety_constraint] trainer = ConstitutionalDPOTrainer(model, ref_model, constraints=constraints)
Practical Considerations and Best Practices
When implementing DPO for real-world applications, consider the following tips:
a) Data Quality: The quality of your preference data is crucial. Ensure that your dataset:
Covers a diverse range of inputs and desired behaviors
Has consistent and reliable preference annotations
Balances different types of preferences (e.g., factuality, safety, style)
b) Hyperparameter Tuning: While DPO has fewer hyperparameters than RLHF, tuning is still important:
β (beta): Controls the trade-off between preference satisfaction and divergence from the reference model. Start with values around 0.1-0.5.
Learning rate: Use a lower learning rate than standard fine-tuning, typically in the range of 1e-6 to 1e-5.
Batch size: Larger batch sizes (32-128) often work well for preference learning.
c) Iterative Refinement: DPO can be applied iteratively:
Train an initial model using DPO
Generate new responses using the trained model
Collect new preference data on these responses
Retrain using the expanded dataset
Direct Preference Optimization Performance
This image delves into the performance of LLMs like GPT-4 in comparison to human judgments across various training techniques, including Direct Preference Optimization (DPO), Supervised Fine-Tuning (SFT), and Proximal Policy Optimization (PPO). The table reveals that GPT-4’s outputs are increasingly aligned with human preferences, especially in summarization tasks. The level of agreement between GPT-4 and human reviewers demonstrates the model’s ability to generate content that resonates with human evaluators, almost as closely as human-generated content does.
Case Studies and Applications
To illustrate the effectiveness of DPO, let’s look at some real-world applications and some of its variants:
Iterative DPO: Developed by Snorkel (2023), this variant combines rejection sampling with DPO, enabling a more refined selection process for training data. By iterating over multiple rounds of preference sampling, the model is better able to generalize and avoid overfitting to noisy or biased preferences.
IPO (Iterative Preference Optimization): Introduced by Azar et al. (2023), IPO adds a regularization term to prevent overfitting, which is a common issue in preference-based optimization. This extension allows models to maintain a balance between adhering to preferences and preserving generalization capabilities.
KTO (Knowledge Transfer Optimization): A more recent variant from Ethayarajh et al. (2023), KTO dispenses with binary preferences altogether. Instead, it focuses on transferring knowledge from a reference model to the policy model, optimizing for a smoother and more consistent alignment with human values.
Multi-Modal DPO for Cross-Domain Learning by Xu et al. (2024): An approach where DPO is applied across different modalities—text, image, and audio—demonstrating its versatility in aligning models with human preferences across diverse data types. This research highlights the potential of DPO in creating more comprehensive AI systems capable of handling complex, multi-modal tasks.
_*]:min-w-0″ readability=”16″>
Conclusion
Direct Preference Optimization represents a significant advancement in aligning language models with human preferences. Its simplicity, efficiency, and effectiveness make it a powerful tool for researchers and practitioners alike.
By leveraging the power of Direct Preference Optimization and keeping these principles in mind, you can create language models that not only exhibit impressive capabilities but also align closely with human values and intentions.
#2023#2024#agent#agreement#ai#AI optimization techniques#AI preference alignment#AI systems#applications#approach#Artificial Intelligence#audio#Bias#binary#challenge#comparison#comprehensive#content#data#data quality#direct preference#direct preference optimization#domains#DPO#efficiency#extension#Future#GPT#GPT-4#human
0 notes
Text
Conspiratorialism as a material phenomenon

I'll be in TUCSON, AZ from November 8-10: I'm the GUEST OF HONOR at the TUSCON SCIENCE FICTION CONVENTION.
I think it behooves us to be a little skeptical of stories about AI driving people to believe wrong things and commit ugly actions. Not that I like the AI slop that is filling up our social media, but when we look at the ways that AI is harming us, slop is pretty low on the list.
The real AI harms come from the actual things that AI companies sell AI to do. There's the AI gun-detector gadgets that the credulous Mayor Eric Adams put in NYC subways, which led to 2,749 invasive searches and turned up zero guns:
https://www.cbsnews.com/newyork/news/nycs-subway-weapons-detector-pilot-program-ends/
Any time AI is used to predict crime – predictive policing, bail determinations, Child Protective Services red flags – they magnify the biases already present in these systems, and, even worse, they give this bias the veneer of scientific neutrality. This process is called "empiricism-washing," and you know you're experiencing it when you hear some variation on "it's just math, math can't be racist":
https://pluralistic.net/2020/06/23/cryptocidal-maniacs/#phrenology
When AI is used to replace customer service representatives, it systematically defrauds customers, while providing an "accountability sink" that allows the company to disclaim responsibility for the thefts:
https://pluralistic.net/2024/04/23/maximal-plausibility/#reverse-centaurs
When AI is used to perform high-velocity "decision support" that is supposed to inform a "human in the loop," it quickly overwhelms its human overseer, who takes on the role of "moral crumple zone," pressing the "OK" button as fast as they can. This is bad enough when the sacrificial victim is a human overseeing, say, proctoring software that accuses remote students of cheating on their tests:
https://pluralistic.net/2022/02/16/unauthorized-paper/#cheating-anticheat
But it's potentially lethal when the AI is a transcription engine that doctors have to use to feed notes to a data-hungry electronic health record system that is optimized to commit health insurance fraud by seeking out pretenses to "upcode" a patient's treatment. Those AIs are prone to inventing things the doctor never said, inserting them into the record that the doctor is supposed to review, but remember, the only reason the AI is there at all is that the doctor is being asked to do so much paperwork that they don't have time to treat their patients:
https://apnews.com/article/ai-artificial-intelligence-health-business-90020cdf5fa16c79ca2e5b6c4c9bbb14
My point is that "worrying about AI" is a zero-sum game. When we train our fire on the stuff that isn't important to the AI stock swindlers' business-plans (like creating AI slop), we should remember that the AI companies could halt all of that activity and not lose a dime in revenue. By contrast, when we focus on AI applications that do the most direct harm – policing, health, security, customer service – we also focus on the AI applications that make the most money and drive the most investment.
AI hasn't attracted hundreds of billions in investment capital because investors love AI slop. All the money pouring into the system – from investors, from customers, from easily gulled big-city mayors – is chasing things that AI is objectively very bad at and those things also cause much more harm than AI slop. If you want to be a good AI critic, you should devote the majority of your focus to these applications. Sure, they're not as visually arresting, but discrediting them is financially arresting, and that's what really matters.
All that said: AI slop is real, there is a lot of it, and just because it doesn't warrant priority over the stuff AI companies actually sell, it still has cultural significance and is worth considering.
AI slop has turned Facebook into an anaerobic lagoon of botshit, just the laziest, grossest engagement bait, much of it the product of rise-and-grind spammers who avidly consume get rich quick "courses" and then churn out a torrent of "shrimp Jesus" and fake chainsaw sculptures:
https://www.404media.co/email/1cdf7620-2e2f-4450-9cd9-e041f4f0c27f/
For poor engagement farmers in the global south chasing the fractional pennies that Facebook shells out for successful clickbait, the actual content of the slop is beside the point. These spammers aren't necessarily tuned into the psyche of the wealthy-world Facebook users who represent Meta's top monetization subjects. They're just trying everything and doubling down on anything that moves the needle, A/B splitting their way into weird, hyper-optimized, grotesque crap:
https://www.404media.co/facebook-is-being-overrun-with-stolen-ai-generated-images-that-people-think-are-real/
In other words, Facebook's AI spammers are laying out a banquet of arbitrary possibilities, like the letters on a Ouija board, and the Facebook users' clicks and engagement are a collective ideomotor response, moving the algorithm's planchette to the options that tug hardest at our collective delights (or, more often, disgusts).
So, rather than thinking of AI spammers as creating the ideological and aesthetic trends that drive millions of confused Facebook users into condemning, praising, and arguing about surreal botshit, it's more true to say that spammers are discovering these trends within their subjects' collective yearnings and terrors, and then refining them by exploring endlessly ramified variations in search of unsuspected niches.
(If you know anything about AI, this may remind you of something: a Generative Adversarial Network, in which one bot creates variations on a theme, and another bot ranks how closely the variations approach some ideal. In this case, the spammers are the generators and the Facebook users they evince reactions from are the discriminators)
https://en.wikipedia.org/wiki/Generative_adversarial_network
I got to thinking about this today while reading User Mag, Taylor Lorenz's superb newsletter, and her reporting on a new AI slop trend, "My neighbor’s ridiculous reason for egging my car":
https://www.usermag.co/p/my-neighbors-ridiculous-reason-for
The "egging my car" slop consists of endless variations on a story in which the poster (generally a figure of sympathy, canonically a single mother of newborn twins) complains that her awful neighbor threw dozens of eggs at her car to punish her for parking in a way that blocked his elaborate Hallowe'en display. The text is accompanied by an AI-generated image showing a modest family car that has been absolutely plastered with broken eggs, dozens upon dozens of them.
According to Lorenz, variations on this slop are topping very large Facebook discussion forums totalling millions of users, like "Movie Character…,USA Story, Volleyball Women, Top Trends, Love Style, and God Bless." These posts link to SEO sites laden with programmatic advertising.
The funnel goes:
i. Create outrage and hence broad reach;
ii, A small percentage of those who see the post will click through to the SEO site;
iii. A small fraction of those users will click a low-quality ad;
iv. The ad will pay homeopathic sub-pennies to the spammer.
The revenue per user on this kind of scam is next to nothing, so it only works if it can get very broad reach, which is why the spam is so designed for engagement maximization. The more discussion a post generates, the more users Facebook recommends it to.
These are very effective engagement bait. Almost all AI slop gets some free engagement in the form of arguments between users who don't know they're commenting an AI scam and people hectoring them for falling for the scam. This is like the free square in the middle of a bingo card.
Beyond that, there's multivalent outrage: some users are furious about food wastage; others about the poor, victimized "mother" (some users are furious about both). Not only do users get to voice their fury at both of these imaginary sins, they can also argue with one another about whether, say, food wastage even matters when compared to the petty-minded aggression of the "perpetrator." These discussions also offer lots of opportunity for violent fantasies about the bad guy getting a comeuppance, offers to travel to the imaginary AI-generated suburb to dole out a beating, etc. All in all, the spammers behind this tedious fiction have really figured out how to rope in all kinds of users' attention.
Of course, the spammers don't get much from this. There isn't such a thing as an "attention economy." You can't use attention as a unit of account, a medium of exchange or a store of value. Attention – like everything else that you can't build an economy upon, such as cryptocurrency – must be converted to money before it has economic significance. Hence that tooth-achingly trite high-tech neologism, "monetization."
The monetization of attention is very poor, but AI is heavily subsidized or even free (for now), so the largest venture capital and private equity funds in the world are spending billions in public pension money and rich peoples' savings into CO2 plumes, GPUs, and botshit so that a bunch of hustle-culture weirdos in the Pacific Rim can make a few dollars by tricking people into clicking through engagement bait slop – twice.
The slop isn't the point of this, but the slop does have the useful function of making the collective ideomotor response visible and thus providing a peek into our hopes and fears. What does the "egging my car" slop say about the things that we're thinking about?
Lorenz cites Jamie Cohen, a media scholar at CUNY Queens, who points out that subtext of this slop is "fear and distrust in people about their neighbors." Cohen predicts that "the next trend, is going to be stranger and more violent.”
This feels right to me. The corollary of mistrusting your neighbors, of course, is trusting only yourself and your family. Or, as Margaret Thatcher liked to say, "There is no such thing as society. There are individual men and women and there are families."
We are living in the tail end of a 40 year experiment in structuring our world as though "there is no such thing as society." We've gutted our welfare net, shut down or privatized public services, all but abolished solidaristic institutions like unions.
This isn't mere aesthetics: an atomized society is far more hospitable to extreme wealth inequality than one in which we are all in it together. When your power comes from being a "wise consumer" who "votes with your wallet," then all you can do about the climate emergency is buy a different kind of car – you can't build the public transit system that will make cars obsolete.
When you "vote with your wallet" all you can do about animal cruelty and habitat loss is eat less meat. When you "vote with your wallet" all you can do about high drug prices is "shop around for a bargain." When you vote with your wallet, all you can do when your bank forecloses on your home is "choose your next lender more carefully."
Most importantly, when you vote with your wallet, you cast a ballot in an election that the people with the thickest wallets always win. No wonder those people have spent so long teaching us that we can't trust our neighbors, that there is no such thing as society, that we can't have nice things. That there is no alternative.
The commercial surveillance industry really wants you to believe that they're good at convincing people of things, because that's a good way to sell advertising. But claims of mind-control are pretty goddamned improbable – everyone who ever claimed to have managed the trick was lying, from Rasputin to MK-ULTRA:
https://pluralistic.net/HowToDestroySurveillanceCapitalism
Rather than seeing these platforms as convincing people of things, we should understand them as discovering and reinforcing the ideology that people have been driven to by material conditions. Platforms like Facebook show us to one another, let us form groups that can imperfectly fill in for the solidarity we're desperate for after 40 years of "no such thing as society."
The most interesting thing about "egging my car" slop is that it reveals that so many of us are convinced of two contradictory things: first, that everyone else is a monster who will turn on you for the pettiest of reasons; and second, that we're all the kind of people who would stick up for the victims of those monsters.
Tor Books as just published two new, free LITTLE BROTHER stories: VIGILANT, about creepy surveillance in distance education; and SPILL, about oil pipelines and indigenous landback.

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/2024/10/29/hobbesian-slop/#cui-bono
Image: Cryteria (modified) https://commons.wikimedia.org/wiki/File:HAL9000.svg
CC BY 3.0 https://creativecommons.org/licenses/by/3.0/deed.en
#pluralistic#taylor lorenz#conspiratorialism#conspiracy fantasy#mind control#a paradise built in hell#solnit#ai slop#ai#disinformation#materialism#doppelganger#naomi klein
308 notes
·
View notes
Text
Just Maybe
Prompt:
18) You have a blind date planned with a member of your bias group. When that member is running late, your bias steps in.
Pairing: Stray Kids Bang Chan x Reader
Genre: Fluff
Song rec as you read: That's Okay - D.O
.
You arrived fifteen minutes early because you always do.
Something could be said for the small comfort of not having to rush into what would be an unknown situation.
You had picked the Italian restaurant you were currently sitting in for much of the same reason. Sure, the food wasn't exactly fantastic, but you had been there before, and it was casual enough to make it seem like you weren't trying too hard (while being nice enough to show you were taking this date seriously.)
When you arrived, the waiter had led you to a quaint table near the window, clearly meant for a couple who intended to talk past dinner and into dessert. Fifteen minutes ago, you had thought it could be that kind of night as well.
Hyunjin had agreed to meet you at 7PM. By his reasoning, he didn't have much of a schedule that day, so he should have no issue making it work. Admittedly, you didn't know what exactly his schedule entailed, seeing as how your best friend, Min had set most of this up. Hyunjin was a mystery to you, just a name attached to a phone number.
At 7:03, you smoothed your napkin against your thigh and tried to remind yourself: It's only three minutes. People are late. It's not personal.
By 7:12, you had read the menu three times without actually digesting a single word of it.
By 7:23, you got the first text.
Hyunjin: running late. parking is a nightmare! 10 min out
Ten minutes was nothing. You had waited twenty already.
The waiter came by yet again, always polite, not pushy. "Can I get you something other than water?"
You looked down at the glass that already had condensation rings staining the table, a silent signal of how long you had sat. Trying to keep your emotions in check, you blinked up before addressing him.
Pasting on a smile that you knew would not appear natural you managed. "I'm alright for now. Thank you."
By 7:41, you were all but crushed. The quiet sting of embarrassment threatened to cause your tear ducts to mutiny. You couldn't even be angry, you were just...deflated.
For a moment, you found yourself cursing your own optimism. You hadn't really minded the idea of a blind date. After all, how were you expected to randomly stumble upon your soul mate in the wild? Something about this all had its appeal and to see it blow up in your face...
It's not like you felt either way about Hyunjin. You'd never met him. Min had set you up, swearing the two of you would click instantly.
The first step to that, of course, was to have him actually show up.
Tapping your phone on the table, you contemplated whether you actually wanted to engage in texting him again or just call it a night.
"Excuse me," a voice broke through the spiral of your thoughts. "Is this seat taken?"
Your breath caught before your mind could actually register the person in front of you. It wasn't Hyunjin.
Quickly flipping through your mental filing cabinet of people you had met, you landed on an answer.
His name was Chan.
You had only met once or twice before. He was another of Min's friends. While Hyunjin had generally remained elusive, Chan often showed up for any housewarming or party your friend put together. He seemed kind, reliable, someone she probably should have thought of before sidling you with a "free spirit."
And now he was here?
Tilting your head, words failed you as you took him in. He was dressed relatively casually, in a t-shirt with long sleeves underneath. He wore light washed jeans and sneakers. His outfit was low effort and casual, giving his handsome features the ability to really shine through.
You blinked. "I-I'm waiting for someone?"
You hadn't meant for the end of that sentence to tilt up at the end into a question, but you supposed it was still valid. You didn't really know if you were waiting for someone anymore.
"Hyunjin?" he smiled knowingly, sliding into the seat before you.
"How did you know?" you asked, trying to swallow down the embarrassment even further.
Chan held up his phone, his smile growing into a grimace. "Min texted me. She said you were here...and I quote..." He glanced to his phone screen. ""They're being too polite to leave, but they sound like they want to throw themselves into traffic.""
Your eyes grew wide as you looked from his phone to your own. Min had definitely sent you several texts to check in, but you hadn't expected for her to call in back up.
You couldn't help it as a rasp of a laugh escaped from your lips. "She's not wrong. I was actually about to leave, and rage order some delivery to get to my apartment before I did."
Chan lifted his brows. "I completely understand if you'd like to go home to just decompress and be alone...however, I do have another offer."
Narrowing your eyes, you waited for him to continue.
"You can eat your feelings in the comfort of your own home OR join me for something that is not this place."
Looking around you, you loosened a sigh. You found it amusing that Chan had been sitting here less than five minutes and could already pick up just how badly you wanted to leave. A simple restaurant that had been your go to for specific situations was now tainted. Even if it wasn't a terribly big deal, you wouldn't help but think of it as "that place you got stood up at" going forward.
"What did you have in mind?"
Chan's grin filled his face again. "7/11, sidewalk seating, and if you play your cards right, I may even spring for banana milk."
You considered him, then the cold breadbasket in front of you, and lastly your phone.
Ding.
Hyunjin: ugh. something came up. mind if we reschedule? i'm soooo sorry
You stared at the message for a moment, your mind blank with irritation. Tapping the lock button, you slipped your phone into your bag.
Standing, you nodded. "Lead the way."
.
The summer air wrapped around you like a warm scarf as you walked side-by-side with Chan, the glow of storefronts and late-night coffee shops painting his face in a rainbow of color. The city was buzzing with activity as people spilled out of restaurants and laughter floated above the hum of traffic. The sights and sounds were oddly comforting, making you feel like you were a small part of a living thing.
Side eying you without saying a word, Chan reached down. Intertwining your fingers, he smiled to himself before looking up again.
It was now your turn to look at the hand that now encapsulated yours.
Well...alright.
You tried to remind yourself to remain cool. He wasn't as much of a stranger as Hyunjin would have been, and you wouldn't have been surprised for him to hold your hand had he been here. What about this was so different?
"So did Min dial you in as my back up date?"
"Like she has a litany of men on speed dial just for the occasion?"
You felt your cheeks grow hot.
"She was complaining to me about how Hyunjin flaked," Chan smiled, still not looking at you directly. "And I happened to be in the area."
"You decided to show up all on your own?"
The resulting squeeze to your hand was enough of an answer.
A comfortable silence descended on the two of you, but it wasn't long until you reached your destination.
"After you," Chan hummed, holding the convenience store door open.
Inside, the store was fluorescent and familiar. Bright packaging, pastel drinks, and a small, refrigerated section with perfectly symmetrical food trays stacked and ready for the taking. Unsure of where to start, you idled toward the cooler, hovering in front of your choices.
Chan reached in immediately, grabbing two sets of sushi without hesitation.
"You've done this before," you smirked.
"Oh, absolutely," he nodded. "This is date number...seven, I think?"
You opened your mouth before closing it and trying again. "Seven?"
It was now his turn to smile. "With myself."
You laughed, throwing caution to the wind and grabbing a tray at random. You followed Chan to the register where he added two cans of peach sparkling water and a pack of peanut snacks "for dessert."
Trailing him outside, Chan led you to a quiet side street tucked beside a bakery that had closed for the night. There was a little patch of grass, a low concrete step, and a view of the bustling street adjacent to you. Chan sat first, offering his hand to help lower you down. Plopping unceremoniously beside him, you watched as he carefully spread out the food as if on a picnic.
He cracked open your drink and passed it over. "Sorry, I know it's not necessarily what you had planned for the night."
You tapped your drink lightly against his. "I should be the one apologizing to you."
"For what?" he asked. "Coming with me to do exactly what I was planning to do anyhow?"
You rolled your eyes. "For making you feel bad enough to include me."
"No," he smiled softly, his eyes lingering on you before he finally looked down to his sushi. "I'm happy you gave me the chance"
The first bite was better than it had any right to be. The rice was somehow still slightly warm, the salmon buttery, and the soy sauce packets not overly salty. You both ate in silence for a moment, heads bent like kids in a cafeteria until Chan glanced back up at you.
"You don't have to answer this," he started. "But were you actually excited about the date with Hyunjin? Or was this a "sure, why not" situation?"
You chewed slowly, considering. "Honestly? Somewhere in between. I wasn't expecting any immediate fireworks, but I thought...maybe it'd be nice."
Chan nodded. "I've agreed to dates for the promise of less."
"Yeah?" you asked, furrowing your brows.
"You see," he said quietly. "The problem with having low self-esteem is that you agree to almost anyone who propositions you. I once agreed to a date where the person showed up to the cafe with their mom. They were in their late 20's."
"I'm sorry, what?" you gasped, trying not to choke on your food.
He shook his head. "I panic ate an entire croissant just to avoid eye contact."
"No, no, no, not that," you chuckled. "But yes, that does suck. What do you mean you have low self-esteem? You're gorgeous."
Lifting a brow he gave you an incredulous look. "Self-esteem isn't always measured by your outward appearance."
You chewed on your lip, unsure if your comment was shallow or not.
Chan heaved a sigh. "Not to get existential or self-deprecating, but I have a lot of thoughts on myself, how I interact with others, and the way I generally conduct my life. It's a lot to reconcile when facing something that seems as easy as dating."
You nodded, beginning to understand. "You're scared to show your true self to someone, just to be turned away."
"Ding, ding, ding," he chuckled. "We have a winner."
Taking a deep breath to bolster your confidence, you reached over and set your hand on his knee. Giving a small squeeze, you chanced direct eye contact. "Winner in more ways than one."
Chan's smile was slow, but it eventually took over before he looked away, back at his food.
There was a pause, but just like any before, still pleasant.
Chan set his tray aside and leaned back on his hands, straightening his legs. "If I continue being honest, will I scare you away?"
You huffed a laugh before setting your food aside as well. Something in his tone had shifted, so you weren't exactly sure what was coming.
"I take silence as a yes," he nodded before it turned into a head shake. "But only in this one situation. Silence is never consent-"
"Chan!" you blurted. "While appreciated that you recognize that, please continue."
He glanced up to the sky, seeming to search the stars for the words he hadn't yet said.
"I've kind of been hoping I would run into you again," he said quietly. "After Min's housewarming...I couldn't get you out of my head. You were wearing this green sweater and making silly little jokes to calm her down about her lack of a turn out."
"You remembered what I had on?" you asked, mouth quirking into a half smile.
Chan shrugged. "It was a nice sweater."
The world seemed to narrow a little, not in a claustrophobic way, but in that soft, quiet tunnel you get when something real is about to start. You reached out, stealing a piece of sushi from his tray mostly to ground yourself.
"You know," you said, chewing. "This could have been a really crappy night."
Chan nodded, turning his attentions from the sky back to you. "Could've been for both of us."
"But now..." you trailed off.
He tilted his head. "Now?"
You met his gaze and smiled slowly. "Now it feels like a real date."
Chan watched you for a beat before leaning forward just enough so that you could meet him halfway if you wanted to.
Oh, and you wanted to.
The kiss was gentle, unhurried. It wasn't showy or dramatic, just a moment of warmth pressed into your mouth, like something shared in a secret. His hand grazed your knee, tentative but not unsure, and when he pulled back, he stayed close enough that you could feel his breath against your cheek.
"Okay," he whispered. "Maybe I owe Hyunjin a thank you."
You laughed against his shoulder. "Maybe I owe him one too."
The night stretched out before you, wide and open and impossibly full. Somewhere between being stood up and a convenience store picnic, something wonderful had happened. It wasn't what you expected, but maybe, it was what you'd been waiting for all along.
#bang chan#chan bang#chan x reader#stray kids#skz#skz x reader#skz x you#chan fanfic#chan scenario#stray kids x reader#stray kids x y/n#stray kids x you#hyungin#chan x you#chan x y/n
79 notes
·
View notes