#it could be misunderstood or it’s not coherent enough

[jayskai]

i want to write a comic about being trans but ughhh why is it so hard ,,,

[mymreaderlibrary]

Could always go for some old man yaoi between Price and an older veteran reader so here are some ramblings.

[Old friends, pre relationship, yearning, beard buddies, ramblings/ no coherency].

[Length: 739 words]

|

|

|

The 141's mission calls for a specialization that they don't meet (at least not well enough to risk failure), but what would you know Price has an old buddy who can help. The thing is he's a veteran who's not all that interested in exiting his retirement especially considering he lost a limb. However he's not one to abandon those in need and with a little convincing he decides to assist, offering his expertise mostly off of the field.

In their downtime y/n and Price get to catch up and spend some time together over drinks. It's been a while since they've seen each other considering how differently they now lead their lives. Y/n is embarrassed to admit he's already getting grey hairs as Price questions his brutally short buzzcut. He never used to cut his hair like that even in his younger years. Price jokes that he's disappointed that y/n is trying to hide his new silver fox look. Y/n says he'll get to see it when it starts growing into his beard. Maybe it's the alcohol but the teasing lingers in both of their minds, feeling more sincere than joking.

Even if they've spent a lot of time apart there's still things they know about each other by heart. Particularly how apprehensive y/n is to being assisted with anything. There's no fragility allowed as he'd rather grin and bare it than accept help, especially with his new(ish) disability. Price doesn't speak on the matter, but he does get frustrated at the other man's stubbornness, a subtle crinkle to his eye as he watches his friend struggle.

Speaking of stubbornness, the 141 get to "enjoy" having two hardheaded older men on the same team. Double the scolding and double the exasperated sighing, somehow it's like being together has combined them into the ultimate grumpy old man. Laswell wouldn't admit it but once their voices start to layer over each other she tunes them out. They're gonna be going in circles for a while and it's near impossible to stop them. It would be almost comedic if it didn't happen all the damn time. ( Worst of all is that most of their arguments are them agreeing on something, but thinking the other misunderstood due to how they phrased their response).

Night times are often nostalgic, neither of them are heavy sleepers so if one wakes up so does the other. They try to get as much sleep in when they can, but if they feel like reminiscing over times long gone then who's to stop them. Stories about Price's younger years in training and y/n's experience with that one damn drill instructor who seemed to hate his guts. Stories about how shitty their first long range shots were and how they could barely even hit the target. Stories about all the times they found themselves in some form of trouble be it on the field or in base. Y/n remembers the time they had to share one shitty cot that they almost punched each other over. Price remembers when they were stuck in a tundra for over a month and were so excited to have a hot shower again they just went at the same time. Neither of them mention how much they actually enjoyed that closeness and how both of them were too cowardly to initiate anything. But they do mention the day they got split up, sent to entirely different countries. Y/n admits to being disappointed he hadn't gotten the chance to say goodbye... Price says he's glad they didn't, sometimes goodbye's like those are bad luck.

They take a moment to look at each other, see scars old and new, wrinkles forming in places that used to be so smooth. Price can't help but look at y/n's eyes, seeing that familiar color that never seemed to fade despite all its seen. He thinks back to that warm shower over a decade ago and how they looked then too. The steam that rose from y/n's chest, the subtle brush of skin they shared as they tried to dance around each other. He remembers that expectation of something more that never came.

They squeeze each other's hands, but the moment passes in silence and neither of them move. Eventually they break eye contact and drift back to sleep.

It seems even now they’re still too cowardly.

[paintedstarz]

Heiji Hattori's lack of boundary was a lot for most people to handle, but for Hakuba, and his lack of understanding of body language and tones, does not mind. Hakuba took it as a privileged offering of comfort Hattori had opposed on him.

This was untrue, as the two had never naturally gotten along. Although, through mingling upon cases and with a enough of Shinichi's exasperated sighs, the two learned to tolerate one another.

Hakuba felt delighted to have Hattori offer supposed kindness, "a step to ease the tension in our new acquaintance," he thought aloud one day, striking a bizarre look from Hattori.

"Nah, I do that wit' everyone. It bothers m'real bad though, but I don' mean to," Heiji rambled, though the quality of his explanation was so poor that Hakuba squinted in confusion as if he could attempt to see it better.

"Ya don' seem ta' mind though. Ya surprised me!" Hattori smiled. He jutted out his elbow to prod at Hakuba's side playfully.

Hakuba looked down in a bashful manner, feeling a pleasant warmth blossom in his chest. No, he didn't mind. No one else had ever been so outward at expressing their affections like Hattori had been towards him.

"People like Kudoh' fer' example hate it, then again he's always pushin' everyone away," Heiji murmured, frowning. His gaze suddenly fixated elsewhere, as if his trail of thought were drifting off up into the sky.

Hakuba blinked, he had to constantly remind himself of Kudo Shinichi, who had suddenly reappeared. Shinichi never stuck around in one place, but that never stopped Heiji from taking almost every chance to mention him.

"Kazuha doesn't like it either, but I forget that a lot." Heiji chuckles, erupting a small, unexpected smile from Hakuba. He hadn't meant to smile, it was just instinctive to smile whenever Hattori did.

"Ran is nice 'bout it, but ya can't hover 'round a girl yanno?  Especially since we ain't datin' or nothin'." Heiji clicked his tongue disapprovingly, leaning over the fencing, before resting his elbows over the rusted edge.

"Ya met Ran before, yeah?" Heiji cocks his head to the side, looking over at Hakuba.

Hakuba has the slightest instinct to frown, but he fought off the urge and pursed his lips instead.

"You're interested in Mouri's daughter?" Hakuba asked, voice slightly out of it's usual pitch.

"What?" Hattori's shoulders tense.

Hattori opened his mouth slowly, exhaling through his nose deeply as if Hakuba had uttered the stupidest words in coherent human language.

"We ain't datin' so it'd be weird for me ta' invade her personal space, and it'd look weird too! I ain't know her all dat' well!" Heiji spat, he seemed rather revolted at the thought.

"She's pretty n'all but I ain't got my eye on her, yanno?" Heiji said, straining his jaw in a rather exaggerated pout.

Hakuba meekly lowered his shoulders and drove his hands deeper into his pockets as if they were to swallow him.

I want to disappear, he thought miserably.

He had misunderstood as usual.

Hattori smacked the back of his head playfully, "Don' get tha' wrong idea pal," he warned.

What on earth did that mean?

His reaction was so defensive. Did that insinuate he's dating someone else?

Who did Hattori have eye for? Hakuba frowned.

Timidly, Hakuba's eyes wander to make  contact with Hattori's. Hattori beamed a playful smile in response, glowing in the golden sun. The light danced over his dark skin, engulfing him in the golden light as if it were honey.

It was quite a beautiful color, the sunny hues over such melanin-rich skin.

Hakuba raises a hand to wipe at his face in disbelief.

He frowns deeply at the thought; Have I caught eye for Hattori? 

[dropout-if]

"please don’t leave me. " for Travis?

From this ask game!

You were eighteen— life was much easier in a very naive way. At least for you, freshly accepted into Stanford, the pride and heart of the neighborhood. You were just eighteen and already considered an example.

High schoolers never worried about your achievements the same way their parents did, but an excuse was an excuse: Downtown Heights was big enough that people outside the neighborhood knew about it, yet small and poor enough that you going to Stanford is an event important enough to get your name on the local newspaper (‘The one who made it out,’ they called you ‘Local Legend’). And such an excuse was enough to garner a party on your name.

Some called it a goodbye. And you were one of them.

Amidst the vibrant lights and thumping music of the senior year party, you found yourself in a quieter corner of the backyard, away from the frenetic energy of the crowd. The night sky was dotted with stars, the air alive with laughter and youthful exuberance.

"please don’t leave me. " for Travis?

Travis, usually reserved and composed, stood next to you, swaying slightly as he tried to maintain his balance. The faint smell of alcohol wafted through the air, it was clear that he’d had a few drinks—though he promised Uma he wouldn’t, it seemed like a couple of beers would be the only reason he’d stand the party in the first place. Travis’ normally serious expression had softened, you could even spot a glint of mischief in his usually solemn eyes.

“You know,” Travis slurred a little—cleared his throat as soon as he noticed his words weren’t as precise as always, “I never thought I’d see you at one of these things.”

Amused by the contrast between his usual demeanor and his state at the moment, you chuckled, “You’re one to talk.”

“Why would I not come? Uma isn’t here.”

“You hate people,” you replied instantly “And Uma is sick.”

Travis rolled his eyes a little, mind a little too hazy to answer coherently. His gaze fixed on you, eyes narrowing slightly, studying you.

“You’ve changed,” he murmured, voice carrying a very distinct hint of nostalgia.

A breeze swept through the backyard, ruffling your hair, you shiver slightly in the cool night air. Travis noticed—spent a few moments thinking and thinking, that’s Travis for you, always thinking, never acting—and then you saw him fumbling with the oversized jacket he was wearing. His intention was clear. Despite his poor coordination, Travis intended to put it around your shoulders.

“Here,” he said, “You can’t get a cold before you leave.”

As Travis stood by your side—swaying slightly—as you wrapped yourself in his jacket, you were allowed to see such a rare and uncharacteristic display of vulnerability from him. Almost as if the little alcohol Travis had consumed had managed to peel away at his layers, chipping at his guardedness, revealing a side of your best friend you’d seldom seen prior to that night.

You frowned, said gently, “Travis,” there was an edge to your voice that you didn’t like. You tried again, “Are you— are you okay?”

“No.”

Travis winced, looked away briefly before he met your gaze again. There was a moment of silence, and then he spoke, words raw and unfiltered, “[Name]— Please don’t leave me.”

His out-of-place earnestness—his words, his eyes—caught you off guard. Rendered you silent, practically feeling the weight of his words hanging in the air—in your shoulders, too—a mixture of uncertainty and what you might have misunderstood as longing.

“I’m here, Travis. I’ll always be.”

Your best friend snorted, leaned back against the wall. Travis didn’t look at you when he called you a “Liar.”

[k-s-morgan]

What do you think of the movie Joker from 2019?

Ask 2: And what would you change about it if you could (movie Joker)

-------

Ooh. I have two very contradictory opinions on it.

On the one hand, I loved it. It has some really powerful moments; the music is amazing; I appreciated the link with Bruce, weird and unlikely as it was; I loved seeing the Joker's dance, his increasing theatricality, him painting a bloody grin on his face and briefly leading the crowd.

On the other hand, I don't think this movie should exist. At least not in the Batman universe and definitely not in relation to Batman's Joker. Because Arthur Fleck is not the Joker and he never will be.

Age is a shallow but still relevant detail. Young Joaquin Phoenix would have made an amazing Joker, but he's too old to be pre-Batman Joker now. What's the age difference between him and Bruce, 30 years? This makes their canon decades-long dynamic impossible to achieve because by the time Bruce becomes Batman, Arthur Fleck is going to be 55 or so. Their brutal physical fights, the roof jumping, the flirting, none of it can realistically happen with this kind of age difference.

Then there is the character himself. I can buy this background: I'm sure Joker experienced abuse in his past, and taking Killing Joke, he tends to see his pre-rebirth version as someone pathetic and misunderstood. So the story of Arthur Fleck works - to an extent. The problem is, in Killing Joke, we get just several short glimpses, but in a full-blown movie, they had to show the essence of Joker properly. Joker is exceptionally intelligent. He's great at chemistry, he's talented with explosives, he's fluent in several languages and he has a very high IQ. Despite being thin and seemingly fragile, he's flexible, agile, and fierce in combat. Arthur Fleck is none of these things. Well, okay, he showed impressive speed when he killed Randall, but it's not enough. He's an average older man with a disorder who's been facing abuse for his entire life. He's not a genius, he doesn't come up with any complex many-layered schemes even after transforming into Joker. He's not chaotic and he doesn't seem interested in anarchy - he just goes with the flow.

Also, motivations. Joker doesn't have a motivation, that's the whole point. He's obsessed with Batman and chaos. Arthur goes after people who hurt him, which makes him a very different kind of character. He also gives off some incel vibes, with the way he conjures a fantasy about his neighbor and feels entitled to break into her apartment on its basis.

Joker might have undergone a serious transformation after his one bad day, but he didn't become a completely different person. That's the essence of Killing Joke: you don't become someone entirely new after trauma. I don't see almost anything from Joker in Arthur and I don't believe for one second that he can turn into him.

If I had a chance to make a Joker movie, I would have expanded on the multiple-choice past idea. I'd show different bits from different scenarios, linking them organically into one almost-coherent plot and leaving Bruce's imprints all over each one. He'd be one unchanging constant in every version of the past Joker thinks he might have had because he plays a central role in shaping his current identity.

[abyssalcunters]

CHAPTER 3 THE BIRD FUCKIN GETS IT

Irene, when she had first arrived on the landship, was the subject of much interest. An Iberian Inquisitor, resigning her title to join Rhodes? One associated with the mysterious, misunderstood Skadi and Specter, no less? For a brief, overwhelming period she was the talk of the town. 

Not infrequently during that time, she was asked every thinkable variation of “Is Skadi really like a walking Catastrophe?” by almost every other resident of Rhodes. It confused her! Sure, Skadi was inhumanly strong, but so were the other Hunters. She, and they, were still mostly human, just a little more, right? Nothing to compare to world-shaping forces of nature.

So she thought.

read on ao3

It dawns on her, now, as Skadi braces one hand beneath her knee and effortlessly folds Irene in half, that there might be something to those stories. 

She bucks her hips unconsciously as Skadi’s teeth find her neck again, blunter than Laurentina’s but no less deadly feeling as she lines Irene’s carotid with the barest shred of restraint.

Briefly, Irene wonders how it would feel if she bit down. She knows, logically, that there would be no way she could survive that, but some deep, hidden away facet of her throbs at the idea.

She closes her eyes as Skadi rocks their hips together, as Laurentina squeezes a breast just tight enough to sting, and sees them both painted in blood.

Her hips twitch. She whines, low and desperate, as Skadi pulls away from her throat to nose at her ear. 

“Do you want,” she murmurs, breath short, a guttural, primal quality to it that sends a full-body shiver down Irene’s spine, “do you want me here?”

She brushes her knuckle over Irene’s hole and she jolts, hips twitching upwards into unsatisfying nothing, keening pitifully at the denial. She does; she wants, desperately, but all her wrung-out mind can think to do is to try and pull Skadi closer. To bare her throat to her Hunter, and let herself be hunted, but trying to move Skadi is like trying to will gravity to lessen its pull.

A sharp tug at her hair brings her briefly back to coherence, shivering as Laurentina traces a finger along her jaw. 

“She asked you a question, my pretty little bird,” she says, fingers firm in Irene’s hair as her other hand, precise with long practice, squeezes just under the corners of her jaw, “Answer her.”

Irene’s moan hitches in her throat, pitching an octave up from something lustful to desperate as Skadi brushes against her again, and nods as furiously as she can against Laurentina’s fist in her hair.

Laurentina’s grip tightens, and Irene’s heart leaps into her throat.

“Words, birdie.”

“Please,” Irene chokes out, and the dam breaks. “Please, Skadi I want you I need you to touch me there, I-” she sucks a breath between her teeth as Skadi’s finger brushes, presses more firmly, then retreats once more, the repeated denial welling tears at the corners of her eyes, “I need you in me Skadi please I need you to- to fuck me, please-”

In a blink, Skadi is on her again, slick fingers rubbing tight circles, wrenching a sob of torture relieved from Irene’s throat as she kisses away the tears rolling down her cheeks. “Okay, little bird,” she murmurs, “Okay.” 

Skadi works her open with tender expertise, one finger gauging, massaging, coaxing choked whines; two pushing deeper, twisting, receding to do it again. She curls her fingers with each thrust, milking full-throated moans along with dribbles of pre that Laurentina wastes no time in spreading back down over Irene’s clit, thumb brushing insistent just under her tip.

Irene melts. 

She sinks deep, both Hunters hands on her, in her, sinks into the haze of submission, of being taken care of, into the anticipation that Skadi will fuck her soon, Irene sinks. There is safety to be found here, of all places; bruised, bleeding slightly, breathless and dizzy with arousal and with an absolutely destroyed pair of panties, Irene finds safety in- or on, in one case -the hands of those who not long ago would have instead been at the end of her blade.

But here she is theirs, and there’s nothing anyone can do about it but her. Between two dangerous, vicious, destructive, violent, deadly, wholeheartedly loving people she finds control, and gives it to them willingly. Irene offers her heart to them with open palms, and trusts them to do right with it.

Skadi does, by wedging one terribly strong hand beneath Irene’s hip and bodily flipping her onto her stomach.

Laurentina laughs that tinkling, musical laugh of hers, tossing her panties aside before settling, legs spread, right in front of where Irene’s face is pressed to the mattress, “Oh, Orca, you really do have the best fucking ideas sometimes, dont you?” 

Skadi’s satisfied smile is tangible against Irene’s hip as she kisses over it, hands trailing up Irene’s thighs drawing soft hums of appreciation before a sharp, pitched gasp as she hooks her thumbs under the crease of Irene’s cheeks and spreads her, gentle but inexorable. 

Irene whines, embarrassed at being so exposed but more so at how she can feel herself dripping onto the sheets, each minute shift of Skadi’s fingers rippling needy twitches over her hips. Then Skadi dips her head, and drags her tongue hot and wet against her, and Irene’s arms give out. 

She licks again, circling, tip of her tongue just dipping into Irene’s stretched out hole, and Irene whines brokenly as Laurentina tugs her head up to rest on her thigh. 

“Don’t you think she’s waited long enough, Orca?” 

“Mmm,” Skadi hums in lieu of response, pressing one last kiss to Irene’s twitching ass before sitting back on her heels, “Sorry, little bird… you looked delicious.”

Irene can only pant into Laurentina’s lap, whining as Skadi fingers more cool lube into her, not conscious or caring of how she’s drooling a little, or how when Skadi’s fingers brush just so inside her again she twitches and leaks on the sheets. 

Skadi pulls away, briefly, and Irene half-sobs at the emptiness, clenching around nothing, digging her fingers into the meat of Laurentina’s thigh, but then Skadi’s clean hand is tangling itself in her hair and she’s pressing the head of her cock up against her and Irene whimpers, involuntary, trying to push herself back onto her but Skadi’s grip on her hip is iron, holding Irene exactly where she wants her as she rocks forward, gentle, easing her into it.

Breath catching in her throat, Irene tries to say something, anything, to beg for more for faster for anything more than this agonizingly careful pace. It comes out as whimpers, mumbled, incoherent against Laurentina’s skin, but through some divine providence (or perhaps Skadi just being really good at sex) she squeezes Irene’s hip once and snaps her hips forward, controlled, and hilts herself fully in her. 

Irene’s whimpers turn to moans, full, deep, throaty, and as Skadi sets her pace Laurentina’s hand replaces Skadi’s in her hair and tugs her head firmly between her legs. Irene gasps, the tug sending static racing down her spine, hands coming up from where they were uselessly gripping at the sheets to hold Laurentina’s hips as she kisses unrefined, desperate, hungry at her folds. 

Laurentina hums in appreciation, meeting Skadi’s eyes for a moment. “Seems like our- mm -our birdie here is hungry for more, hmm?” Her fingers tighten in Irene’s hair, forcing her to look up, “Should we fuck you senseless, darling?” 

Irene nods, furiously, and Laurentina laughs, sharp and hungry, and yanks her face closer by the hair, hips rising to grind, hard, against Irene’s tongue. 

Skadi bends forward to mouth at the nape of Irene’s neck, breasts pressed against her back, giving her a moment to adjust to Laurentina’s face-fucking.

Then she draws back, leaving Irene whimpering into Laurentina’s pussy at the loss, before pounding back into her, demanding, wringing the pleasure from Irene’s desperate, pliant body, quiet huffs of breath against her neck in time with Irene’s sobs of pleasure at each savage thrust.

It’s all Irene can do to keep her mouth open for Laurentina, head swimming with the ease she’s handling her, at the fact that she’s being useful- that she’s being used -that her body can be the vessel for these women’s pleasure. Laurentina rocks against her mouth with abandon, spreading slick over Irene’s lips, tongue, cheeks, chin, the tip of her nose bumping against her clit on each upstroke and Laurentina sings her delight, unabashed and bright. 

Skadi is quieter in her intensity, hot breath in hissed exhales washing over Irene’s hypersensitive, painted neck in a scorching balm as her fingers dig new bruises into her hips. Like Irene has noticed before on the battlefield, she is single-minded in her focus, determined. 

Unstoppable. 

Irene’s whole body shakes as Skadi thrusts into her, hands on her hips enough to almost wrap around her whole waist, canting her hips up to fuck her tip against Irene’s prostate with each unforgiving thrust, driving her up and up and up and she’s almost there, her vision fuzzing white at the edges, sparks dancing down her spine, and her legs give out. 

She sobs, agonized, she’s so close, but Skadi doesn’t miss a beat in hooking hands up under the crease of her thighs and lifting Irene’s ass back where she needs it, the angle allowing her to somehow hit deeper, strokes over Irene’s g-spot winding white-hot need though her gut, dragging screaming moans out of her, babbling nonsensical, incoherent pleas in Iberian into Laurentina’s crotch, prayers not to her God but a greater power. 

Then Irene’s world splits, and she screams, orgasm coursing through her like a tsunami, convulsing in their arms- or trying to. The Hunters’ hands remain firm, even as Irene shakes apart between them they continue, Skadi fucking her, thrusts growing shaky as Irene clenches around her, Laurentina’s cries peaking as she grinds deeper into Irene’s already dripping mouth.

Distantly, through the staticy haze of overstimulation, the overwhelming supernova of heat bursting through her, Irene registers someone crying. 

Then Laurentina’s thighs clamp around her face, pressing against wet tear tracks as she keens in her ecstasy, the lack of air and Laurentina’s heady scent blurring her senses into one unbelievable concentration as Skadi, still holding her off her useless legs, digs her teeth into Irene’s shoulder, snaps her hips deep, spilling hot bliss into her, and Irene comes undone.

She comes around to the murmur of quiet Ægirean; someone had shifted her so she was sitting across Laurentina’s lap, head resting delicately in the crook of her shoulder. She blinks away the haze lingering in her mind, opening bleary, slightly unfocused eyes to see the bright red of two otherworldly predators staring back at her. 

Laurentina smiles, softer than any Irene has seen from her before, the corners of her eyes crinkling as she looks down at this fragile woman in her lap. “Welcome back, Irene.” 

Her throat is killing her. Irene taps her head tiredly against Laurentina’s shoulder in lieu of response. 

Skadi smiles, slight, tender, brushing her fingers through Irene’s hair, soothing her aching scalp. “I will prepare the bath.” Irene blinks. “To clean,” Skadi clarifies, as if Irene didn’t understand, then: “Like Shark said. We want to take care of you.” She presses a kiss to Irene’s head before standing, and Irene tries to ignore how her eyes prickle at the sentiment.

“Look at her,” Laurentina says, shaking her head in mock exasperation, “drawing a bath? She’s gone so soft, here on land. Then again,” she glances down at Irene, meeting her eyes, “some things here are worth all the other bullshit.”

Then Laurentina leans down to kiss her, gently, fingers weaving into her hair. Irene comes up for air, still a bit short of breath, resting her forehead against Laurentina’s neck. 

She’s filthy; face mostly coated in Laurentina’s slick, Skadi’s cum trickling out of her and down both of her legs. She’s battered; bruises painted over her throat and shoulders, slowly drying blood laced across her collar and over her breasts, hips and neck and everywhere in between sore from being so thoroughly used. She’s exhausted, wrung out physically and mentally, and she still can’t move her legs. 

“Yeah,” she says, the lump in her throat and welling at the corner of her eyes squeezing her voice into a whisper, “it really is worth it.” 

Irene, former Inquisitor, now simply a woman, for the first time she can remember, says “It’s worth it” and means it.

[mariacallous]

With OpenAI’s ChatGPT now a constant presence both on social media and in the news, generative artificial intelligence (AI) models have taken hold of the public’s imagination. Policymakers have taken note too, with statements from Members addressing risks and AI-generated text read on the floor of the House of Representatives. While they are still emerging technologies, generative AI models have been around long enough to consider what we know now, and what regulatory interventions might best tackle both legitimate commercial use and malicious use.

What are generative AI models?

ChatGPT is just one of a new generation of generative models—its fame is a result of how accessible it is to the public, not necessarily its extraordinary function. Other examples include text generation models like DeepMind’s Sparrow and the collaborative open-science model Bloom; image generation models such as StabilityAI’s Stable Diffusion and OpenAI’s DALL-E 2; as well as audio-generating models like Microsoft’s VALL-E and Google’s MusicLM.

While any algorithm can generate output, generative AI systems are typically thought of as those which focus on aesthetically pleasing imagery, compelling text, or coherent audio outputs. These are different goals than more traditional AI systems, which often try to estimate a specific number or choose between a set of options.  More traditional AI systems might identify which advertisement would lead to the highest chance that an individual will click on it. Generative AI is different—it is instead doing its best to match aesthetic patterns in its underlying data to create convincing content.

In all forms (e.g., text, imagery, and audio), generative AI is attempting to match the style and appearance of its underlying data. Modern approaches have advanced incredibly fast in this capacity—leading to compelling text in many languages, cohesive imagery in many artistic styles, and synthetic audio that can impersonate individual voices or produce pleasant music.

Yet, this impressive mimicry is not the same as comprehension. A study of DALL-E 2 found that it could generate images that correctly matched prompts using the word “on” just over one quarter of the time. Other basic spatial connections (such as “under” and “in”) led to even worse results. ChatGPT shows similar problems. As it is merely designed to string words together in a likely order, it still cannot reliably pass basic tests of comprehension. As is well documented by Professor Gary Marcus, ChatGPT may often fail to “count to four… do one-digit arithmetic in the context of simple word problem… figure out the order of events in a story… [and] it couldn’t reason about the physical world.”

Further, text generation models constantly make things up—OpenAI CEO Sam Altman has said as much, noting “it’s a mistake to be relying on [ChatGPT] for anything important right now.” The lesson is that writing convincing, authoritative-sounding text based on everything written on the internet has turned out to be an easier problem to solve than teaching AI to know much about the world. However, this significant shortcoming did not stop Microsoft from rolling out a version of OpenAI’s technology for some users of its search engine.

Still, this sense of authenticity will make generative AI appealing for malicious use where the truth is less important than the message it advances, such as disinformation campaigns and online harassment. It is also why an early commercial application of generative AI is to create marketing content, where the strict accuracy of the writing simply isn’t very important. However, when the media website CNET started using generative models for writing financial articles, where the truth is quite important, the articles were discovered to have many errors. These two examples offer a glimpse into two separate sources of risk from generative AI— commercial applications and malicious use—which warrant separate consideration, and likely, distinct policy interventions.[1]

Handling the Commercial Risks of Generative AI

The first category of risks comes from the commercial application of generative AI. Many companies want to use generative AI for various business applications that are far more general than simply generating content. For the most part, generative AI models tend to be especially large and relatively powerful, and so while they may be particularly good at generated text or images, they can be adapted for a wide variety of tasks.[2]

The most prominent example may be Copilot, an adaptation of OpenAI’s GPT-3. Developed by GitHub, Copilot integrates GPT-3 into a more specific tool for generating code, aiming to ease certain programming tasks. Other examples include the expansion of image-generating AI in helping to design video game environments and the company Alpha Cephei, which takes open-source AI models for speech analysis and further develops them into enterprise voice recognition products. The key concern with collaborative deployment using generative AI is that neither company may sufficiently understand the function of the final AI system.[3]

The original developer solely developed the generative AI model but cannot see the full extent to how it is used when it is adapted for another purpose. Then a “downstream developer,” which did not participate in the original model development, may adapt the model and integrate its outputs into a broader software system. Neither entity has complete control or a comprehensive view into the whole system. This may increase the likelihood of errors and unexpected behavior, especially since many downstream developers may overestimate the capacity of the generative AI model. This joint development process may be fine for processes where errors are not especially important (e.g., clothing recommendations) or where there is a human reviewing the result (e.g., a writing assistant).

However, if these trends extend into generative AI systems used for impactful socioeconomic decisions, such as educational access, hiring, financial services access, or healthcare, it should be carefully scrutinized by policymakers. The stakes for persons affected by these decisions can be very high, and policymakers should take note that AI systems developed or deployed by multiple entities may pose a higher degree of risk. Already, applications such as KeeperTax, which fine-tunes OpenAI models to evaluate tax statements to find tax-deductible expenses, are raising the stakes. This high-stakes category also includes DoNotPay, a company dubiously claiming to offer automated legal advice based on OpenAI models.

Further, if generative AI developers are uncertain if their models should be used for such impactful applications, they should clearly say so and restrict those questionable usages in their terms of service. In the future, if these applications are allowed, generative AI companies should work proactively to share information with downstream developers, such as operational and testing results, so that they can be used more appropriately. The best-case scenario may be that the developer shares the model itself, enabling the downstream developer to test it without restrictions. A middle-ground approach would be for generative AI developers to expand the available functionality for, and reduce or remove the cost of, thorough AI testing and evaluation.

Information sharing may mitigate the risks of multi-organizational AI development, but it would only be part of the solution. This approach to help downstream developers responsibly leverage generative AI tools only really works if the final system is itself regulated, as will be the case in the EU under the AI Act, and as is advocated for in the U.S.’s Blueprint for an AI Bill of Rights.

Mitigating Malicious Use of Generative AI

The second category of harm arises from the malicious use of generative AI. Generative models can create non-consensual pornography and aid in the process of automating hate speech, targeted harassment, or disinformation. These models have also already started to enable more convincing scams, in one instance helping fraudsters mimic a CEO’s voice in order to obtain a $240,000 wire transfer. Most of these challenges are not new in digital ecosystems, but the proliferation of generative AI is likely to worsen them all.

Since these harms result from malicious use by scammers, anonymous harassers, foreign non-state actors, or hostile governments, it may also be much more challenging to prevent them, compared to commercial harms. However, it might be reasonable to require a certain degree of risk management, especially by commercial operations that deploy and profit from these cutting-edge models.

This might include tech companies that provide these models over API (e.g., OpenAI, Stability AI), through cloud services (e.g., the Amazon, Google, and Microsoft clouds), or possibly even through Software-as-a-Service providers (e.g., Adobe Photoshop). These businesses control several levers that might partially prevent malicious use of their AI models. This includes interventions with the input data, the model architecture, review of model outputs, monitoring users during deployment, and post-hoc detection of generated content. Manipulating the input data before model development is an impactful way to influence the resulting generative AI, because these models greatly reflect that underlying data. For example, OpenAI uses human reviewers to detect and remove “images depicting graphic violence and sexual content” from the training data for DALL-E 2. The work of these human reviewers was used to build a smaller AI model that was used to detect images that OpenAI didn’t want to include in its training data, thus improving the impact of the human reviewers. The same type of model can also be used at other stages to further prevent malicious use, by checking to see if any images submitted by users, or the images generated by generative AI, might contain graphic violence or sexual content. Generally, the practice of using a combination of human reviewers and AI tools for removing harmful content may be an effective, if not sufficient, intervention.[4] The development of generative models also may provide an opportunity for intervention, although this research is just emerging. For example, by getting iterative feedback from humans, generative language models can become moderately more truthful, as suggested by new research from DeepMind.[5]

User monitoring is another tactic that may bear fruit. First, a generative AI company can set transparent limits on user behavior through the Terms of Service. For instance, OpenAI says its tools may not be used to infringe or misappropriate any person’s rights, and further limits some categories of images and text that users are allowed to generate. OpenAI appears to have some system to implement these terms of service, such as by denying obvious requests for harassing comments or statements on famous conspiracy theories. However, one analysis found that ChatGPT responded with misleading claims 80% of the time, when presented with a catalog of misinformation narratives. Going further, generative AI companies could monitor users, using algorithmic tools to flag requests that may suggest malicious or banned use, and then suspend users who become repeat offenders.

In a more nascent approach, researchers have proposed using patterns in generated text to identify it later as having come from a generative model, or so-called watermarking. However, it is too early to determine how such a detection might work once there are many available language models, available in different versions, that individual users are allowed to update and adapt. This approach may simply not adapt well as these models become more common.

Collectively, these interventions and others might add up to a moderately effective risk management system. However, it is highly unlikely it would be anywhere near perfect, and motivated malicious actors will find ways to circumvent these defenses. In general, the efficacy of these efforts should be considered more like content moderation, where even the best systems only prevent some proportion of banned content.

It is still the early days of generative AI policy

The challenges posed by generative AI, both through malicious use and commercial use, are in some ways relatively recent, and the best policies are not obvious. It is not even clear that “generative AI” is the right category to focus on, rather than including individually focusing on language, imagery, and audio models. Generative AI developers could contribute to the policy discussion by disclosing more specific details on how they develop generative AI, such as through model cards, and also explain how they are currently approaching risk management.

It also warrants mention that, while these harms are not trivial, there are more pressing areas in which the U.S. needs AI governance, such as protections from algorithms used in key socioeconomic decisions, developing meaningful online platform policy, and even passing data privacy legislation. If perhaps not a priority, it is worth considering regulations for commercial developers of the largest AI models, such as generative AI.[6] As discussed, this might include information sharing obligations to reduce commercialization risks, as well as requiring risk management systems to mitigate malicious use. Neither intervention is a panacea, but they are reasonable requirements for these companies which might improve their net social impact. This combination might represent one path forward for the EU, which was recently considering how to regulate generative models (under the distinct, but related term, “general-purpose AI”) in its proposed AI Act.[7] This would raise many key questions, such as how to enforce these rules and what to do about their considerable international impact. In any case, if the EU or other governments do take this approach, it is worth keeping policies flexible into the future, as there is still much to be learned about how to mitigate risks of generative AI.

[surrealsunday]

Crazy theory anon here.

Thanks for your response because talking with others is also helping me make sense of things in my own head. I like what you said and I agree that there might be another thing that breaks KP up again after they survive the current situation because I keep waffling between what I already wrote and this theory. I wonder if it might have anything to do with the thumb drive. Tawan showing up has put the spotlight on Porsche because they’re trying to frame him as a scapegoat or, at the very least, a distraction but I find it interesting that they keep bringing up his parents death. So much that it can’t be written off as insignificant.

I don’t necessarily think this is what’s gonna happen but I’m curious if some connection between Porsche and Kinns family might arise when they look at whatever is on that drive. The one thing that is in Porsches favor is that the mole has been there long before he and Kinn even met but if there’s a connection to the family since he was a child? That could be misconstrued in a myriad of ways and place doubt in the minds of the people preaching his innocence, especially if the family actually is connected to their deaths in some way. That might make it seem like some form of motive for Porsche and further serve as a misdirect from whoever is orchestrating this whole thing. It could also potentially drive an even bigger wedge between KP. On the flip side, if Porsche finds out that, say Korn for example, was responsible for literally destroying his life and causing lifelong trauma to him and his brother and then forcing him into this life that he fought against so hard? That could also be really bad and maybe cause Porsche to doubt Kinn for the first time since they’ve been together. I mean, even before Porsche, I think the main goal was to keep Kinn isolated and paranoid and now that can’t happen if he and Porsche are solid. He’s no longer a sitting duck in the middle of this mafia pond.

Like, I find it interesting that when Vegas toying with Porsche was obviously not working and only really bringing KP closer together, they throw a wrench into it with Tawan showing up who just so coincidentally was held captive by the Italians. It’s just so obviously suspect.

I don’t actually think that Porsche will take off with Vegas tbh I just kinda like the idea but I never even considered Kinn walking away from the pool because he trusted Porsche to handle himself. That’s so interesting! I think the truth is also that Porsche just doesn’t have the option to tell Vegas to back off and like he said in the past to Kinn, he has to play his role when it comes to that situation “manners, you know?” Because as much as Porsche is a little shit who’s goal in life is to be an endearing pain in Kinns ass, he understand his place with the other members of the family for the most part. I think at the very least Porsche is on guard with Vegas but might be smart enough to act like he’s buying this “I’m so misunderstood” schtick and still seem as passive as Vegas thinks he is. However deep it runs, I just don’t think Porsche is trusting of Vegas anymore but he knows he needs to keep making him think he’s creating a weak link with his tactics. If that makes sense? That’s why hearing people in the fandom say things like “Porsche needs to stay away from Vegas” is so frustrating because he doesn’t really have the option to do that anymore than he has the option to stay away from anyone else in the family because they’re the goddamn mafia. He doesn’t have the same leeway he has with Kinn and I hope Kinn is finally getting over his jealousy enough to understand that now.

This is a little out there I admit but like I said, I loooooove speculating.

Thanks for your response though, you’re very good at articulating your thoughts and it kinda helps me with all this information bouncing around my head. I struggle to get it all out in a coherent way sometimes so I very much enjoy interacting with you.

Ohhhhh I'm so glad you came back. This is fun!

And god you're totally right. There is the thumb drive and that whole thing about Porsche's parents to contend with too. I'm super intrigued to see how all of it will come together. You raise some really interesting thoughts which reminds me of something I had originally considered happening (but lost sight of in the drama of this ep). I also had thought Korn's (maybe) involvement in Porsche's parents deaths could potentially throw a curveball and cause the final hurdle before Kinn and Porsche form a united front. My speculation (and disclaimer... truly based on nothing but what I could imagine when late night musing) was that maybe whatever the 'truth' is, it's awful enough for Porsche to want out of mafia life. I really can't imagine a situation where he steps away from Kinn - maybe it would be because Porchay has been pulled into the world? - but if he did, it wouldn't be because he didn't love Kinn, but because he had to get away for Porchay's sake. And him leaving would be a betrayal of sorts to Kinn? I dunno. I still can't totally see Kinn pulling a gun on him like he does in the preview under those circumstances. But I think all those lose threads have to come together to cause some drama. That's for sure.

"I think the main goal was to keep Kinn isolated and paranoid and now that can’t happen if he and Porsche are solid. He’s no longer a sitting duck in the middle of this mafia pond." I so totally agree. Tawan sees Porsche (and Kinn's soft heart which he's given to Porsche) as a weakness. He's wrong though. Porsche makes Kinn stronger - and vice versa. And that is a lesson I can't wait for Tawan to learn. Vegas too sees Porsche as an opportunity to weaken and destabilize Kinn. But he's just going about it in a way that would actually work (if it weren't for the fact that Kinn and Porsche are so devoted to one another). Vegas knows if he took Porsche from Kinn, Kinn would be totally lost. And I agree... it's way too suspect that Tawan shows up right when Vegas's plan so dramatically fails. Vegas has a long-game plan for sure, and I would be shocked if that didn't intersect with Tawan in some way. And perhaps Tawan was inserted into things because Kinn and Porsche were proving themselves to be too devoted to one another - so additional measures were taken.

You actually sort of changed my mind on the whole 'Porsche running with Vegas' thing and I am seriously considering it now. Maybe it just won't last long? And it will be in the interest of Porsche clearing his name? Now that you've raised the possibility with me, I am considering it more.

And OMG we are on the EXACT SAME PAGE when it comes to your thoughts about Porsche's interactions with Vegas. Yes! I totally get why people get frustrated watching and want Porsche to tell Vegas off or just leave or whatever but... and maybe this is just because I relate to Porsche a fair bit... I think I would react the exact same way as Porsche? The sort of 'play nice' thing while being on guard and no longer trusting Vegas. It's like when Vegas kisses Porsche and Porsche does a sort of meek move trying to get away (vs. the way he responds with Kinn, for eg). That made so much sense to me. He can't piss off Vegas and reject him or tell him off. Not without the very real danger that comes with that (which he acknowledged at the end of ep 3 in recognizing that Kinn saved his life when threatened by the minor family). Porsche can tell off Kinn all he likes - he always could - and do so without fearing for his life. There was always a level of trust there. That doesn't extend to Vegas and it especially doesn't extend to Vegas now. And of course there is another layer to their interaction at the pool in this ep. Porsche is vulnerable and he sees Vegas being vulnerable in return (yes it's a ploy but Vegas fronts well). He's wary and reserved, but there is a comradery he allows in that moment because, quite frankly, he's feeling a bit lost. It makes sense to me in the context of everything we see happen in the ep. And honestly it makes me super curious to get a better sense of what Porsche is thinking when it comes to Vegas, which I imagine we will get next ep.

And truly it's my pleasure to chat with you 💞💞💞. Your thoughts are extremely well laid out and really make me think! Trust me when I say you are beyond coherent (I hope I am offering the same despite this being the end of a very long work day for me lol). I can't wait to find out what we might be right or wrong about!

[jamdoughnutmagician]

10 characters, 10 fandoms, 10 tags

The lovely @bookshelfdreams tagged me (thanks for the tag!) to do this, and while I can't guarantee this will be a coherent list of characters (and reasons why I like them!) hopefully this will help at least make for interesting reading, if not some very un-hinged thoughts at the very least 🤷‍♀️

Rose Tyler (Doctor Who)

Rose was one of my earliest memories of being obsessed with a character. I started watching Doctor Who in about 2006/during season 2 (I would have been about 8/maybe 9 years old at the time) so I missed her initial introduction in season 1, but I went back and rewatched her stuff with the ninth doctor and I just absolutely loved how she was just an ordinary working class girl who wanted to go on adventures (of course I didn't understand that this was the reason why I loved her so much as a kid but yeah..)

Bucky Barnes (MCU)

Look, I know we all have our opinions/ and criticisms on the mcu (which I totally understand) but there was just something about Bucky that I gravitated towards (Probably because I found Sebastian Stan attractive, but I digress.) A young boy who was willingly wanting to go fight in the war to defend his country and then ending up going through arguably the worst kind of trauma that someone could go through. Then to see his journey and story progress through the years, for me, has been a real treat to watch.

Eddie Munson (Stranger Things)

Eddie became a very quick favourite of mine in such a short space of time (partly to Joseph Quinn playing him brilliantly) but I just loved how his character COULD have been this dark, brooding, misunderstood teen, but he was ACTUALLY a very sweet and warm character, despite what everyone else thought of him. He was just a funky little metal-head D&D nerd, and I loved him for that. I just really need him back for series 5 Please!!

Orla McCool (Derry Girls)

Again, Orla was just someone I gravitated towards, I can't quite put my finger on why, I just really liked her giddy optimism and strange but sweet outlook on things!

Harley Quinn (DCEU) is that what it's called sorry idk about these things! 🤦‍♀️

Margot Robbie's Harley Quinn is very special to me (Especially in her post Suicide Squad era) because she's outlandish and wacky, but also incredibly smart and quick. I'd read a handful of Harley comics and I just think she's really cool.

Elphaba Thropp (Wicked The Musical)

I can't really speak too much about her novel version because I haven't read enough of the book to form an opinion, but stage Elphaba has my entire heart. She's unapologetic in standing up for what she believes in and fighting what she believes is right, and I admire that about her.

Mabel Mora (Only Murders in the Building)

More so I love her relationship between Oliver and Charles, but Mabel's willingness to get stuck in and try to help to solve murders alongside her two slightly older friends was a very interesting dynamic to watch! Also I would kill to have her outfits because girls got style.

Gina Linetti (Brooklyn Nine-Nine)

How could you not love the absolute self-confidence and bravado of Gina Linetti! She was as will forever be the most <3

Tecna (Winx Club)

Although I sat through both seasons of that bad neflix adaptation, nothing compares to the original cartoon in my eyes (at least the first 4 seasons anyway.) I think it has something to do with how it captured my attention as a kid because the idea of pretty fairies to me as a child was very appealing! I think (although the whole group had their own merits) Tecna was my favourite even just from a character design standpoint.

Becky Lynch (WWE)

I don't know if this technically counts since she's a real person, and considering I haven't watched WWE for a while but Becky's in-ring persona, was always a fascinating and exciting watch for me! Especially finding her feet as 'Becky-Two-Belts' and her being 'The Man', she was strong, powerful and for a long time under-rated as a performer, but I loved her so so much!

Tagging (sorry I'm not tagging 10 people, also don't feel like you have to do this at all!)

@the-odd-aardvadillo @unlimiteddream90 @lucyferchrist @ndj-blogs @smute @pherre @amidalasweets @theravensdreamer

[riversebb]

I had a friend who said 'I smile when I touch a life, and laugh when it touches back'. There are only individuals. And all of those individuals live on the same rock, flying through space, at a million miles an hour. Our life spans are short and we are each very small. but if you ever can know what the first sentence means, really experience it, then you will know that the size and duration of a life are of little consequence. you will know an absolute truth without having to name it. and you will be misunderstood if ever you try to financially profit from such a knowing. love is not financially profitable. it could be, but it is not. coherent and capable human relations don't make good advertisements for guns or bombs. this is all there is to be done. understand this thing, and live it as wholly as you can. on the shoulders of those that came before, like the ants that form themselves into a raft in the downpour, we are uplifted by the relations that we choose to participate in. do what you can do, do what is enough for you.

[kyrodo]

Trolling me doesn't earn you any fucking respect from me. When you can show me you can act like a fucking adult, then things will definitely improve.

I sincerely hope that'll be the last time someone says "answer the fucking question" to me without actually asking anything or even making it obvious what they were asking. Are you a fucking idiot? I understand you like to do everything by force but get the fuck away from me. Now that you've seen everything there is to see, who falls under what description now bitch? We talking about me or are we talking about you!? I can count my list of offenses on one hand and not a single one of them were major, can you say the same with a straight face you stupid fuck?

Any time you do anything it's ok and anything anyone else does is a crime against humanity, it doesn't fucking work that way. I'm more than happy to put every unforgivable thing you do out in the open and unlike you everything I post doesn't come out of fucking nowhere. You exploit every little thing that could be misunderstood, but for you, I understand everything. You act radicalized but in actuality you're just an opportunist waiting for the chance to pretend someone else did something. I spend so much time on my posts trying to foolproof it and you still find a way to act like a fucking moron. Somebody will turn on their high beams and you'll just close your eyes and say it's still too Fucking dark. I've had about enough of your bullshit.

Now that I've stopped adopting the way you speak, am I coherent enough for you asshole? You got anything else you want to take out of context, or are you fucking good?

As for Ronnie telling me "I need to stop doing this, I don't know why you're focusing me." We don't need to say what we mean was literally one of the first things you shared with me. Never once did you come clean about your side of things and you always omit your entire history whenever you directly interact with me or try to keep your wording as ambiguous or multi meaning'd as possible. What the fuck did you want me to do? I physically deleted mine, cause I was mad or there were personal thoughts I was always pushing out that I didn't want to leave up. But you initiated the facade and you strong armed it the entire way through, I didn't. I protested it because it kept confusing me and I kept wondering if you were all these things because you kept it up well past when it was no longer meaningful.

There isn't a single person vrc or otherwise with your kind of special needs shit, there is not single person I've met as overly discreet as you. I will never be the one to send someone through fucking bootcamp before I give them even a scrap of trust. Because I will never be like you.

Every cell in my body even now wants to make sure no one ever experiences what I did. What you put me through. If you ever wanted me in your life, there's not a single thing you did that would've made that possible.

I am so sick of feeling constantly confused and uncertain about everything, that I will always go out of my way to ensure everyone feels certain.

[usagiverse]

*whimpering and rolling on the floor, with incomprehensible words*

Too much information to process!! ASADFJKLFIEJNNASJFWER!

Vesper! Sir! was I rude to ask for pronouns and not say mine? *shuddering* sometimes I'm just unintentionally rude *gasp* right! they\them …I am gender fluid but I like they\them more… and we are team green?! I am INFP!

sorry for the inconsistency I'm just so... * happy jumping*…I'll need some time to process... well, everything *shaking their hands* we're going to have so many shenanigans, and fun! and I could never hate you! you are so... is easy for me to identify with you... you're team green, you know how hard it is just to communicate, plus you're dense! you know how is live with that! I'll stick with my webs here

I have so many ideas for pranks for the turtles, and I read all the info! our bunnies need therapy, and you have evil plans, and I want to see them in action…just agh! there's so much to say and my brain is asdadadjkasfj

...anyway you don't need to respond to this, when I calm down could ask a question with more coherence... do you think I'm cute? fr? not creepy or uncomfortable? my brain isn't processing that!

You're not rude for not sharing, Bunnon. You are a mystery, one that does not require being solved. It's not unintentionally rude, but I do see how you might think so. For a long time, I would be "unintentionally rude". All of my questions had an edge. My tone was blunt and unfiltered. I had a hard time understanding things because I was dense, and didn't know when it was improper to ask questions or, mostly, how to ask them. I was berated or looked at like a monster for some of the things I'd say, thinking it was genuine curiosity, thinking it was all play, no harm was meant by it, yet people would take offense. The point is, I understand. I understand it all. And if you, or anyone really, is ever afraid to ask a question or slip in a small note to my inbox, fearful of being misunderstood or shamed, I'm letting you know: don't be. I have been through enough when it comes to communication, and I am still not perfect at it. I still have social anxiety, I still can be blunt and dense, but I can try. I can try to understand, and often times, I think I really do. I think we could all help each other out, by keeping this an open, welcoming, and positive space. Now, all of that to say.. Bunnon, are you sure we're not the same person?? Team Green, Genderfluid chaos.. where's the hidden camera?? You're joshing me, right?? I might post an episode one preview, but I also have 2 frames left on my personal project.. one will take significantly longer than the other.. but the other is almost done...... but I thrive on torture.. ah the decisions.. my evil plans are soon to take hold, with a juicy chapter cover on the way, maybe some extra artwork, definitely more asks in the meantime while waiting... yes. I can feel the energy, I'm so excited and yet I've never done anything like this before so I hope you guys can match my excitement too ! (And yes, you seem to be the adorable type. In a non-creepy way,, in a, "aw, they're so sweet" little gremlin climbing on the walls kind of cute way ? I'm sorry if that made you uncomfortable, I just really like how you type everything as if you're talking, it's very familiar and easy, for me.)

[codasylph]

this is gonna come off as blunt but i mean it kindly; you don't have to have any feeling at all about how a stranger identifies. it's genuinely healthier if you don't. you don't have to try to figure out whether someone is allowed to use certain terms or not, you don't have to try to figure out who's being genuine, it is not your problem. you can just enjoy the idea of "gender: merry trickster". you will literally never have enough context to make a judgment about a complete stranger's sexuality or orientation that isn't making a LOT of assumptions. so if she didn't say anything harmful, what ARE we critiquing? the completely unknown potential she could be lying to the interviewer about feeling not fully straight? the possibility she may have misunderstood herself? straight people beginning to publicly express their sexuality in fun silly ways too? at best it's throwing punches at shadows; at worst it's discouraging people from exploring. you just don't have to stress over stuff like this.

Did you ask similar questions of the people who just straight up said her statements were bi erasure, "straight people bullshit," or something to that effect? And if not, are you reaching out to me because my statement that I felt conflicted gave you hope that you could help me become unconflicted about it? Or did I just write a giant wall of text that did a better job of drawing your attention? Or did my actual reblog somehow come across your dash and you weren't just looking in the notes on someone else's reblog in which case *how*? I have 1 follower and it's a bot. On a more serious note: I *don't* disagree with you. I, in fact, already said in my first reblog that normalizing straight people exploring their gender identity and sexuality would be a very good thing. But I also use my blog to get my thoughts out sometimes and seeing that made me feel conflicted and I wanted to ramble semi-coherently about it.

And you are correct that I don't *have* to have feelings about how a total stranger identifies, but it's also true that I cannot control the fact that, in this case, I do have feelings about it. I feel conflicted about that, too, actually. (Like, I am allowed to have feelings but also Natasha Lyonne does not owe me shit and I do not wish to give the impression that I think she does.)

[education30and40blog]

Early thoughts on regulating generative AI like ChatGPT

See on Scoop.it - Education 2.0 & 3.0

Generative artificial intelligence, like ChatGPT, is a misunderstood emerging technology which poses a number of risks in its potential commercial application and malicious use, including its lack of truthfulness and potential for deception.

With OpenAI’s ChatGPT now a constant presence both on social media and in the news, generative artificial intelligence (AI) models have taken hold of the public’s imagination. Policymakers have taken note too, with statements from Members addressing risks and AI-generated text read on the floor of the House of Representatives. While they are still emerging technologies, generative AI models have been around long enough to consider what we know now, and what regulatory interventions might best tackle both legitimate commercial use and malicious use.

Alex Engler

Fellow - Governance Studies, Center for Technology Innovation

@AlexCEngler

WHAT ARE GENERATIVE AI MODELS?

ChatGPT is just one of a new generation of generative models—its fame is a result of how accessible it is to the public, not necessarily its extraordinary function. Other examples include text generation models like DeepMind’s Sparrow and the collaborative open-science model Bloom; image generation models such as StabilityAI’s Stable Diffusion and OpenAI’s DALL-E 2; as well as audio-generating models like Microsoft’s VALL-E and Google’s MusicLM.

While any algorithm can generate output, generative AI systems are typically thought of as those which focus on aesthetically pleasing imagery, compelling text, or coherent audio outputs. These are different goals than more traditional AI systems, which often try to estimate a specific number or choose between a set of options.  More traditional AI systems might identify which advertisement would lead to the highest chance that an individual will click on it. Generative AI is different—it is instead doing its best to match aesthetic patterns in its underlying data to create convincing content.

In all forms (e.g., text, imagery, and audio), generative AI is attempting to match the style and appearance of its underlying data. Modern approaches have advanced incredibly fast in this capacity—leading to compelling text in many languages, cohesive imagery in many artistic styles, and synthetic audio that can impersonate individual voices or produce pleasant music.

Yet, this impressive mimicry is not the same as comprehension. A study of DALL-E 2 found that it could generate images that correctly matched prompts using the word “on” just over one quarter of the time. Other basic spatial connections (such as “under” and “in”) led to even worse results. ChatGPT shows similar problems. As it is merely designed to string words together in a likely order, it still cannot reliably pass basic tests of comprehension. As is well documented by Professor Gary Marcus, ChatGPT may often fail to “count to four… do one-digit arithmetic in the context of simple word problem… figure out the order of events in a story… [and] it couldn’t reason about the physical world.”

Further, text generation models constantly make things up—OpenAI CEO Sam Altman has said as much, noting “it’s a mistake to be relying on [ChatGPT] for anything important right now.” The lesson is that writing convincing, authoritative-sounding text based on everything written on the internet has turned out to be an easier problem to solve than teaching AI to know much about the world. However, this significant shortcoming did not stop Microsoft from rolling out a version of OpenAI’s technology for some users of its search engine.

Still, this sense of authenticity will make generative AI appealing for malicious use where the truth is less important than the message it advances, such as disinformation campaigns and online harassment. It is also why an early commercial application of generative AI is to create marketing content, where the strict accuracy of the writing simply isn’t very important. However, when the media website CNET started using generative models for writing financial articles, where the truth is quite important, the articles were discovered to have many errors. These two examples offer a glimpse into two separate sources of risk from generative AI— commercial applications and malicious use—which warrant separate consideration, and likely, distinct policy interventions.[1]

HANDLING THE COMMERCIAL RISKS OF GENERATIVE AI

The first category of risks comes from the commercial application of generative AI. Many companies want to use generative AI for various business applications that are far more general than simply generating content. For the most part, generative AI models tend to be especially large and relatively powerful, and so while they may be particularly good at generated text or images, they can be adapted for a wide variety of tasks.[2]

The most prominent example may be Copilot, an adaptation of OpenAI’s GPT-3. Developed by GitHub, Copilot integrates GPT-3 into a more specific tool for generating code, aiming to ease certain programming tasks. Other examples include the expansion of image-generating AI in helping to design video game environments and the company Alpha Cephei, which takes open-source AI models for speech analysis and further develops them into enterprise voice recognition products. The key concern with collaborative deployment using generative AI is that neither company may sufficiently understand the function of the final AI system.[3]

The original developer solely developed the generative AI model but cannot see the full extent to how it is used when it is adapted for another purpose. Then a “downstream developer,” which did not participate in the original model development, may adapt the model and integrate its outputs into a broader software system. Neither entity has complete control or a comprehensive view into the whole system. This may increase the likelihood of errors and unexpected behavior, especially since many downstream developers may overestimate the capacity of the generative AI model. This joint development process may be fine for processes where errors are not especially important (e.g., clothing recommendations) or where there is a human reviewing the result (e.g., a writing assistant).

However, if these trends extend into generative AI systems used for impactful socioeconomic decisions, such as educational access, hiring, financial services access, or healthcare, it should be carefully scrutinized by policymakers. The stakes for persons affected by these decisions can be very high, and policymakers should take note that AI systems developed or deployed by multiple entities may pose a higher degree of risk. Already, applications such as KeeperTax, which fine-tunes OpenAI models to evaluate tax statements to find tax-deductible expenses, are raising the stakes. This high-stakes category also includes DoNotPay, a company dubiously claiming to offer automated legal advice based on OpenAI models.

Further, if generative AI developers are uncertain if their models should be used for such impactful applications, they should clearly say so and restrict those questionable usages in their terms of service. In the future, if these applications are allowed, generative AI companies should work proactively to share information with downstream developers, such as operational and testing results, so that they can be used more appropriately. The best-case scenario may be that the developer shares the model itself, enabling the downstream developer to test it without restrictions. A middle-ground approach would be for generative AI developers to expand the available functionality for, and reduce or remove the cost of, thorough AI testing and evaluation.

Information sharing may mitigate the risks of multi-organizational AI development, but it would only be part of the solution. This approach to help downstream developers responsibly leverage generative AI tools only really works if the final system is itself regulated, as will be the case in the EU under the AI Act, and as is advocated for in the U.S.’s Blueprint for an AI Bill of Rights.

MITIGATING MALICIOUS USE OF GENERATIVE AI

The second category of harm arises from the malicious use of generative AI. Generative models can create non-consensual pornography and aid in the process of automating hate speech, targeted harassment, or disinformation. These models have also already started to enable more convincing scams, in one instance helping fraudsters mimic a CEO’s voice in order to obtain a $240,000 wire transfer. Most of these challenges are not new in digital ecosystems, but the proliferation of generative AI is likely to worsen them all.

Since these harms result from malicious use by scammers, anonymous harassers, foreign non-state actors, or hostile governments, it may also be much more challenging to prevent them, compared to commercial harms. However, it might be reasonable to require a certain degree of risk management, especially by commercial operations that deploy and profit from these cutting-edge models.

This might include tech companies that provide these models over API (e.g., OpenAI, Stability AI), through cloud services (e.g., the Amazon, Google, and Microsoft clouds), or possibly even through Software-as-a-Service providers (e.g., Adobe Photoshop). These businesses control several levers that might partially prevent malicious use of their AI models. This includes interventions with the input data, the model architecture, review of model outputs, monitoring users during deployment, and post-hoc detection of generated content. Manipulating the input data before model development is an impactful way to influence the resulting generative AI, because these models greatly reflect that underlying data. For example, OpenAI uses human reviewers to detect and remove “images depicting graphic violence and sexual content” from the training data for DALL-E 2. The work of these human reviewers was used to build a smaller AI model that was used to detect images that OpenAI didn’t want to include in its training data, thus improving the impact of the human reviewers. The same type of model can also be used at other stages to further prevent malicious use, by checking to see if any images submitted by users, or the images generated by generative AI, might contain graphic violence or sexual content. Generally, the practice of using a combination of human reviewers and AI tools for removing harmful content may be an effective, if not sufficient, intervention.[4] The development of generative models also may provide an opportunity for intervention, although this research is just emerging. For example, by getting iterative feedback from humans, generative language models can become moderately more truthful, as suggested by new research from DeepMind.[5]

User monitoring is another tactic that may bear fruit. First, a generative AI company can set transparent limits on user behavior through the Terms of Service. For instance, OpenAI says its tools may not be used to infringe or misappropriate any person’s rights, and further limits some categories of images and text that users are allowed to generate. OpenAI appears to have some system to implement these terms of service, such as by denying obvious requests for harassing comments or statements on famous conspiracy theories. However, one analysis found that ChatGPT responded with misleading claims 80% of the time, when presented with a catalog of misinformation narratives. Going further, generative AI companies could monitor users, using algorithmic tools to flag requests that may suggest malicious or banned use, and then suspend users who become repeat offenders.

In a more nascent approach, researchers have proposed using patterns in generated text to identify it later as having come from a generative model, or so-called watermarking. However, it is too early to determine how such a detection might work once there are many available language models, available in different versions, that individual users are allowed to update and adapt. This approach may simply not adapt well as these models become more common.

Collectively, these interventions and others might add up to a moderately effective risk management system. However, it is highly unlikely it would be anywhere near perfect, and motivated malicious actors will find ways to circumvent these defenses. In general, the efficacy of these efforts should be considered more like content moderation, where even the best systems only prevent some proportion of banned content.

RELATED CONTENT

TECHTANK

Building guardrails for ChatGPT

Michael J. Ahn and Yu-Che Chen

 Tuesday, February 7, 2023

TECHNOLOGY & INNOVATION

Detecting and mitigating bias in natural language processing

Aylin Caliskan

 Monday, May 10, 2021

TECHTANK

NIST’s AI Risk Management Framework plants a flag in the AI debate

Cameron F. Kerry

 Wednesday, February 15, 2023

IT IS STILL THE EARLY DAYS OF GENERATIVE AI POLICY

The challenges posed by generative AI, both through malicious use and commercial use, are in some ways relatively recent, and the best policies are not obvious. It is not even clear that “generative AI” is the right category to focus on, rather than including individually focusing on language, imagery, and audio models. Generative AI developers could contribute to the policy discussion by disclosing more specific details on how they develop generative AI, such as through model cards, and also explain how they are currently approaching risk management.

It also warrants mention that, while these harms are not trivial, there are more pressing areas in which the U.S. needs AI governance, such as protections from algorithms used in key socioeconomic decisions, developing meaningful online platform policy, and even passing data privacy legislation. If perhaps not a priority, it is worth considering regulations for commercial developers of the largest AI models, such as generative AI.[6] As discussed, this might include information sharing obligations to reduce commercialization risks, as well as requiring risk management systems to mitigate malicious use. Neither intervention is a panacea, but they are reasonable requirements for these companies which might improve their net social impact. This combination might represent one path forward for the EU, which was recently considering how to regulate generative models (under the distinct, but related term, “general-purpose AI”) in its proposed AI Act.[7] This would raise many key questions, such as how to enforce these rules and what to do about their considerable international impact. In any case, if the EU or other governments do take this approach, it is worth keeping policies flexible into the future, as there is still much to be learned about how to mitigate risks of generative AI.

Microsoft provides financial support to the Brookings Institution, including to the Artificial Intelligence and Emerging Technology Initiative and Governance Studies program, where Mr. Engler is a Fellow. Google is a general unrestricted donors to the Brookings Institution. The findings, interpretations, and conclusions posted in this piece are solely those of the author and are not influenced by any donation.

The author acknowledges the research support of CTI’s Mishaela Robison and Xavier Freeman-Edwards.

Footnotes

1. These are two key categories of harms from the use of generative AI, although they are not the only harms. For instance, harms from the development process include copyright infringement (as Getty has charged against Stability AI), undercompensated employees working on potentially harmful data labeling, and the furthering of the business incentive towards massive data collection to fuel ever larger generative models. (Back to top)

2. In other contexts, generative AI even has different names that emphasize its value for re-use. A report from Stanford’s AI community calls them instead “foundation” models and notes in the first sentence that their defining quality is that they “can be adapted to downstream tasks.” (Back to top)

3. The European Union’s proposed AI Act describes this emerging trend, in which multiple entities collaborate to develop an AI system, as the AI Value Chain. It is too early to know how dominant this trend might be, but an enormous increase in venture capital funding suggests a coming expansion of commercial experimentation. (Back to top)

4. However, these companies need to take responsibility for the health and wellness of those human reviewers, who are performing the single most harmful task in the development of a generative AI system. Recent reporting from Time states that Kenyan workers were paid only $2 an hour to categorize disturbing text, and potentially images, on behalf of OpenAI. (Back to top)

5. This is an important research development, but it remains very unclear to what extent large language models will be able to become more routinely and robustly truthful, and they should not yet be assumed to be able to do so. (Back to top)

6. Note that the focus on commercial developers is intentional, and this would not include the open-sourcing of generative AI models, for reasons discussed elsewhere. (Back to top)

7. The malicious use of generative AI poses challenges that are similar to content moderation on online platforms concerning which content should be allowed or disallowed. This makes it an ill-fitting problem for the EU AI Act, which is primarily about the commercial and government use of AI for decision-making and in products. A provision aimed at mitigating generative AI’s malicious use may be a better fit for an amendment to the Digital Services Act, which also has more relevant enforcement provisions. (Back to top)

[newmusicradionetwork]

Andrew Reed & The Liberation “Too Little Too Late”

This is a guy that has gone as DEEP as possible into the experience of LIFE, often isolating himself from society in a rustic cabin surrounded by 13,000 acres of wilderness, to find out for himself, “What’s it all about?” without the influences or hypnotism of a “made up” world, observing Nature and making common cause WITH ITand the way life ACTUALLY works on this planet. Dubbed a “rock mystic” by the Indie press after multiple Billboard and other charting releases, Reed actually lives “liberation” and those in his presence “feel” this Zen or “Unfazed” approach to life. A nonconformist and misunderstood by most, he combines music with intellect in an artistry that flies in the face of the traditional music business world. With him, it is ALL about the artand nothing else. Expressing what cannot be said or communicated any other way. His interviews with press usually leave them confounded, but also inspired. He changes style from one album to the next, and even sees far enough in advance to create, not just a single albums, but album “trilogies.” He will often move on from a budding success before the full commercial potential of a musical direction can be realized, which is scandalous in the music business. He has been known to pull down his entire catalog, like after his Universal Music/INgrooves deal or to further explore material. His live performances are intermittent, but are EPIC and unpredictable, where one doesn’t know what is going to happen. According to Reed, “ROCK was never meant to be pretty, or an over-rehearsed commercial production, but desperate, bombastic and scandalouswhere it could go off the rails at any moment You have to feel like you could throw your life away in the moment” But despite these idiosyncrasies, his work in music & intellectual domains have impact. Not many artists, have the guts to pull all their music downto pursue a particular artistic vision or attainment But that’s the kind of person Andrew is One that can’t be bought or sold. It is all about the musicand one gets that he is literally throwing his life away on itwith a confidence that is inspirational. NEW BOLD RECORDING INNOVATIONS Reed writes not just “concept albums” but “Trilogies” that represent “significant and coherent” bodies of work with unifying sonic themes throughout each trilogy. With musicality pouring out of him, this super guitarist can move from acoustic-oriented to far-out psychedelic rock. Through a “vision” after the catastrophic loss of his son, Reed built a 3 story-mountaintop, World-Class recording studio and compound where he “isolates” and has ventured into bold new recording territory, recording the first of a series of Trilogy albums, 1) As a Bird of the Air 2) As a Lily of the Field 3) As Grass in the Furnace by saturating analog tape running 3x (times) normal speed, capturing performances at the time of inspiration using vintage gear. Also, because of the respect he has gained in the science and health-science communities, especially in the research of consciousness, Reed as incorporated proprietary consciousness-altering technologies into many of his works as well as into the recording process like Monroe Sound Science. “What’s it all about?” Reed’s work serves to help all perhaps create a visionto be completely “liberated” within ourselvesand to live “Guiltless Lives” of complete Self-Acceptance on an odyssey or oscillation of ups and downs, but with an overall momentum UP! Additional Artist/Song Information: Artist Name: Andrew Reed & The Liberation Song Title: Too Little Too Late Publishing: Artists International Broadcasting Publishing Affiliation: BMI Album Title: As A Bird Of The Air Record Label: Artists International Record Label: Artists INTL/WorldSound/INgrooves/UMG Andrew Reed 828-489-9715 areed@multiviewinc.com Radio Promotion: ADD Promotion Gary Lefkowith 203-727-5010 hifiadd@aol.com Manager: WorldSound Warren Wyatt 808-333-4224 warren@worldsound.com Read the full article

[makeste]

BnHA Chapter 328: Pandora’s Box of Discourse

Previously on BnHA: DEKU TOOK A BATH.

Today on BnHA: 

Also Naomasa grew a beard. Goddamn. 

please let this be a cool chapter that plays nice with my ADHD lol

(ETA: lol I feel guilty because a lot of people hated this chapter, but I’m just happy there was a lot of stuff to make fun of, and also that I have another week to work on my backlog of meta posts since the kids were MIA.)

around one month ago?? ah, okay, so we’re gonna find out what was in that Tartarus security file huh

I love that they just randomly set the place on fire

was it necessary to do this in order to escape? no. was it a good idea to set the island they were occupying on fire while they were in the midst of still occupying it? uh. was it cinematic as fuck? fuck yeah

wow it’s a pervert!!

that’s so great that the villains set loose this fine fellow who I’m sure is definitely not a serial rapist. truly the LoV is so noble and misunderstood. they’re just trying to free society from its chains people

oh my god??!

SHANKED!!! oh my god I cheered for Stain before I realized what I was doing. time to have an identity crisis I guess

so he’s all “hey what’s going on.” which, while a respectable question, is something I personally would have waited to ask until I had put a bit of distance between myself and the fiery murder island. but that’s just my personal preference

Stain you really are tenacious I’ll give you that

“what’s the point of escaping prison if you’re not gonna be smart about it” well shit. anyways yeah you’re dead right, society is in the process of collapsing and the outside world is in total chaos, good call there

oh shit

I mean it’s not like we really expecting anything otherwise, but still. fucking brutal. I feel like these guys’ fates were decided the minute that one guy called AFO “scum” back in chapter 94. AFO is unmatched at getting long-term revenge

??

ahh, was it the security footage??

fdsdfk he’s still alive??

and he’s immediately launching into an inappropriately theatrical monologue even as the darkness closes in on him fdlfksjdlk. you know, was it ever confirmed that the other guy back in chapter 297 was Seiji’s dad? I’m just saying

very impressed that he’s still coherent enough to weigh the pros and cons before making the decision to gamble on giving this info to Stain, who at the very least has his own moral code and isn’t allied with AFO. it was definitely still a risk, but as we now know it was also the right call

what a weird alliance. so Stain tells him that he’ll give it to a just person, and the guy is all,

okay for real though I’m gonna need someone to run a DNA test on this guy. maybe it was some kind of cuckold situation?? the other guy had the family resemblance, but this guy absolutely 100% raised Shishikura Seiji and you are not going to convince me otherwise

anyway, so Stain is all,

PRISON GUARD: “???? ??????? what the hell. what the fuck does that fucking mean. I’m dying here, jesus christ, whatever man fuck you”

(ETA: I kind of feel like this might have been Stain’s last appearance in the manga, given all the fanfare. there’s not really much else he can do for the story at this point, and he seems to have gotten all the character development Horikoshi was planning on giving him. so if this really is it, hasta la vista and good riddance I guess.)

DWLFDKSLDK MEANWHILE, OUTSIDE

(ETA: I feel like this is meant to be evocative of that Sermon on the Mount painting, but in a really fucked up way lol.)

if it were me stumbling upon this scene I would just shake my head and walk right back into the flaming building. not getting involved in that mess. sorry not sorry. I’ll take my chances with the fire, especially given that it’s half-assed neutered BnHA fire lol

blah blah blah and so he decided to pass the info on to All Might -- HOT DAMN, HOLY SHIT

NAOMASA HOLY SHIT. THE APOCALYPSE LOOKS GOOD ON YOU, BOY

“I really like that facial scruff thing Aizawa’s got going on, I think I’m gonna get in on that” yes sir. “also thinking of ditching the tie in favor of the bulletproof vest look. also thinking of getting totally fucking jacked.” good lord. except I’m pretty sure that’s just body armor, but also I don’t care. anyway I should probably stop staring and actually read the fucking speech bubbles here lol

“All Might first handed this information over to Nao, and then went to see Deku, and then came back to Nao” thanks for that tidy little summary Horikoshi. we are capable of piecing events together in sequential order, I just want you to know that. but thank you

“so has Deku finally gotten a bath? also, sucks that Stain saved the day, but what are you gonna do” Nao I missed you so fucking much and didn’t even realize. how am I just now realizing that you are the perfect man

for a second I was gonna ask why Tartarus’s security systems would be cut off from the outside world, and then I remembered that’s a basic security control, and then I actually got impressed by how sensible that is. like, it’s been a while since I could genuinely say that the good guys (excluding class 1-A) did something smart. not that it helped them much in the end, but still

anyway so they’re talking about how AFO was able to coordinate the attack by communicating between his horcrux self on the outside and his ugly peanut-faced self on the inside

huh

okay you have my attention. I am taking notes here lol please continue

ah okay so he says that prior to Jakku, the transfer of information between him and his Vestige self was only one-way. but post-Jakku when Deku was in the hospital, he was able to tell what was happening inside the OFA Radical Lisa Frank Dead People Book Club Realm when he touched him. I feel like we established that before, actually. but he didn’t talk about how it actually felt, though

boy we already know this lol. yes AFO can talk with his horcrux self. and he can also communicate with his little bro in OFA too, let’s talk about that sometime why don’t we. what exactly does that imply, based on the rules we’ve established here

my god I cannot get over Naomasa and his fucking facial hair

no wonder All Might was in such a hurry to leave Deku and get back here

like I have no idea what this radio waves nonsense is but my god, people

that jawline. also so it’s a quirk, I see. except last I checked Deku didn’t have a radio waves quirk, so that doesn’t really explain his connection to AFO. but whatever, hopefully we’re at least getting closer to some kind of reveal here

(ETA: since I sometimes forget that other people’s lives don’t revolve around my theory posts, here are the two relevant links if you by chance want to know my thoughts about this.

Hagakure is still The U.A. Traitor™ regardless of whether Deku is passing information on to AFO through his psychic link, which he almost certainly is.

speaking of said psychic link, Deku is a horcrux.

just posting these now, because whenever trippy OFA stuff happens I tend to get an influx of theory asks. so hopefully this will be a bit of a time saver lol.)

-- wait, what

THAT’S what the recording was??!? holy SHIT. I genuinely was not expecting that. y’all wiretapped his fucking telepathy. fucking quirks, man. wild

AND THEY USED THAT POWER TO DETERMINE WHAT WE ALREADY KNEW, HUZZAH. GOOD SHOW

-- oh shit wait lol, except I forgot we’re not talking about 38 days from the present, we’re talking about 38 days from the date the conversation was recorded. heh. um

yeah that’s the face I would make too if All Fucking Might just casually told me we had eight days left until the end times

oh, pardon me. three fucking days

r.i.p. anyone who thought we were going to have another band arc sob. I sure hope Deku is enjoying that nap

(ETA: I realize people were hoping for a longer rest period here, but given that the man warned us all the way back in chapter 306 that we were entering the final act, you can’t really blame him too much when that turns out to be true. anyway but I do recognize that we’ve reached the point in the story where this kind of discourse is going to become a weekly occurrence, simply because there’s no possible way for Horikoshi’s actual endgame to line up perfectly with the variable headcanons of millions of fans, all of whom have wildly differing and in many cases contradictory expectations which can’t possibly all be fulfilled. anyway, so I’m already bracing myself for that lol. this coming year is going to be a wild ride.)

damn, U.A. out here looking like the motherfucking United Nations

-- is this U.A.?? I actually just realized, U.A. is four interconnected buildings, not two. wait holy shit is this Shiketsu?

wait holy SHIT

based on the overwhelmingly powerful vibes of bureaucratic incompetence, I’m thinking this really is the (future) U.N., or whatever organization it is that deals with international hero stuff

“just let them handle it themselves I’m sure they’ll be fine” yeah okay, thanks guys. appreciate it

wait oh shit did he say that it’s not just Japan?

soooo, what you’re telling me is that AFO is this close to bringing about the end of not just Japan, but the entire world, and you guys don’t think it’s a good idea to help the Japanese heroes stop him? so, genuine follow-up question: are you guys already planning your rich people exodus into space a la Wall-E, and that’s why you don’t give a fuck?? like, what??

omg international heroes

these guys are from World Hoodie Mission, right? is this Horikoshi’s way of reminding me to buy tickets

(ETA: and it worked too lol.)

WHO??? WHAT???

don’t tell me you’re introducing yet another badass new female character for me to fall in love with only to watch as you dismember them and/or blow them up, Horikoshi. I’m getting tired of playing this game my dude. don’t lie and tell me this time will be different. we’re not doing this again goddammit

noooooooooooooooooooo

god fucking dammit lmao. [sighs and rips the previous paragraph into shreds]

on behalf of Americans I apologize for our superheroes always being Like This

I also apologize because I love her already and I’m gonna be shameless about it. so fucking shameless you guys

is her fucking hair red white and blue. it is, isn’t it

this is the volume cliffhanger, 100% lol. it will take every ounce of Horikoshi’s willpower not to put her on the volume cover. he’ll have to settle for the spine or the inner cover this time because Deku VS his class 1-a superpals takes precedence. but it will be a close thing let me tell you

tbh it’s that smile that does it for me. she’s definitely All Might’s protege. get out there and show them how it’s done girl. and maybe call Salaam and BRD and see if you can’t convince them to play hooky from their governments as well. why not. world’s ending in three days you guys. “sorry, I’m busy this weekend” ain’t gonna cut it lol

so while I am not fully caught up with Vigilantes, I have read far enough to know that there’s an American hero named Captain Celebrity whose superpower from what I recall is being a humongous douchebag. and while I haven’t read far enough to know what happens to this guy, I can’t say I’m very disappointed to learn that he’s no longer the number one hero in the U.S. (actually, didn’t they kick him out and that’s why he moved to Japan to begin with?). anyway, so my thanks to Horikoshi for having a marginally higher opinion of Americans than Furuhashi, even though we have definitely not done anything to warrant said opinion lately, and you may have inadvertently opened the door to a pandora’s box of discourse lmao

(ETA: lol I went into the tags and they don’t disappoint. “why is she dressed like a flag” because she’s an homage to Captain America and Major Victory and literally every other character on this list. again, I apologize for fictional American superheroes being Like This. “oh boy another thicc waifu to make the fanboys happy” look, tumblr fandom never seems to have a problem thirsting over Dabi or Tomura or Aizawa or Nao, lol, I’m just saying. “where is Captain Celebrity” idk, probably murdered by the exploding bee cartel, let’s just be grateful for our good fortune and try not to Beetlejuice the man.)

anyway, so let’s see if Horikoshi’s recent character development with regards to making Mineta not terrible anymore will apply to other aspects of his writing as well. I know I was making light of discourse just now, but I do think the complaints about him introducing yet another new character at the 11th hour to be cannon fodder in the final battle are absolutely valid. and again, it wouldn’t be a problem if he didn’t keep maiming/killing off his female characters one by one instead of developing them and letting them kick ass long-term. but that said, I will never complain about Horikoshi adding another female character to the series, regardless of how clumsy the attempt may be. go ahead and pander away, just give us more girl power lol

anyway so we’ll see how it goes, but I think I’m gonna be optimistic and let myself hope once again, even though I’m probably gonna regret it lol. it is what it is. she is standing on an airplane just chilling for fuck’s sake. I’m only human. anyway fingers crossed