#AI self-correction

[literaryvein-reblogs]

Do you use AI generators to compile or scrape this information?

no

[aloneeyez]

Some Frenrey comfort doodles I did bcuz yeaaa

[selfship-confession-box]

dude i feel like i selfship with every single ai boy ever like ive got the squip hal9000 edgar AM like whats next 😭🙏

-💕🌀 (heartswirl anon) (i think thats the right emojis 💔)

.

[alanshemper]

absolutely awful take

Let me be clear, it is absolutely impossible to be antisemitic against Israel because Israel is a state, and Zionism is a well defined political project with a history that is documented. It's not antisemitic to question the events and narrative around Oct. 7 and the Supernova Festival, period—enough special pleading for the state of Israel. “Zionism isn’t something special. It’s just nationalism, Jewish nationalism.” False flags are a well known tactic around the world. Refusing to engage with this history and it's impact on the present does not make you a better person or a more noble activist. What happened on Oct 7 was an atrocity, as is the ongoing destruction of Gaza, and that makes it all the more important to understand the facts.

I will say their point about the IOF being in the West Bank at the time (tweet #3) is highly relevant, and something I'd also noted.

[princessclefairyberry]

Oh good yet another reason to unfollow and block half of this godforsaken website.

I swear to fuck the only people on this site are:

• Self-important assholes who think they deserve to tell other people how to spend their pitiful amount of leisure time and money

• Hyper-dedicated fandom, aesthetic, or art blogs who post about absolutely nothing else (this is a compliment)

• Whiny pissbabies who think they are owed other people’s hard work (mostly art) for free

• Blogs specialized in one particular political issue (half are safe to follow, the other half will reveal themselves to be violent bigots as soon as they say anything about a topic other than their usual focus; good luck figuring out which it will be ahead of time!)

• Violent bigots who are aware and open about their bigotry

• Violent bigots who have chugged the koolaid so hard that they think they are morally pure in their bigotry

• Literal teenagers who are so afraid of being shunned that they scream incessantly about what good antis they are, despite not being old enough to remember that the term originated as an insult against people being violent to folks who shipped Reylo. Proship was just Reylo antiship was just anti Reylo it’s all just the fucking Star Wars sequels please get a grip I’m fucikign loisng it over here

• Multi-topic personal blogs that aren’t meant to be curated for other people so good luck (valid, me too)

[jcmarchi]

Reflection 70B : LLM with Self-Correcting Cognition and Leading Performance

New Post has been published on https://thedigitalinsider.com/reflection-70b-llm-with-self-correcting-cognition-and-leading-performance/

Reflection 70B : LLM with Self-Correcting Cognition and Leading Performance

Reflection 70B is an open-source large language model (LLM) developed by HyperWrite. This new model introduces an approach to AI cognition that could reshape how we interact with and rely on AI systems in numerous fields, from language processing to advanced problem-solving.

Leveraging Reflection-Tuning, a groundbreaking technique that allows the model to self-assess and correct its own mistakes in real-time, Reflection 70B has quickly risen to the top, outclassing proprietary models like GPT-4 and Claude 3.5 Sonnet across multiple benchmarks, including MMLU, MATH, and HumanEval.

Reflection 70B is built on the robust Llama 3.1-70B architecture, but its self-refining mechanism sets it apart. Through iterative cycles of reflection, error detection, and output refinement, the model mimics human cognition in an unprecedented way, pushing the boundaries of what AI can achieve. As a result, Reflection 70B offers not only unmatched accuracy but also deeper insights into its decision-making process, a critical feature for applications where transparency and precision are paramount.

What is Reflection 70B

At its core, Reflection 70B is built upon Meta’s open-source Llama 3.1-70B Instruct model. However, what truly sets it apart is its unique ability to engage in a process akin to human reflection—hence its name. This capability stems from a technique called “Reflection-Tuning,” which enables the model to identify and rectify its own errors in real-time, thus improving its accuracy and reliability.

Matt Shumer, CEO of HyperWrite, introduced Reflection 70B with the bold claim that it is “the world’s top open-source AI model.” But what exactly makes this model so special, and how does it stack up against industry giants like GPT-4 and Claude 3.5 Sonnet? Let’s explore.

Understanding Selective Reflection-Tuning: A Paradigm Shift in AI Training

Selective Reflection-Tuning introduces an approach to instruction tuning, where the goal is to improve both the quality of instruction data and its compatibility with the student model being fine-tuned. Traditional methods often focus on improving the data itself but overlook how well the enhanced data pairs align with the learning objectives of the model. Selective Reflection-Tuning bridges this gap by fostering a teacher-student collaboration, where a teacher model introspects on the data and provides refined instruction-response pairs, while the student model evaluates and selects only those improvements that best suit its training needs.

The process consists of two key phases:

Selective Instruction Reflection: The teacher model reflects on the instruction of a given sample and generates a refined instruction-response pair. The student model then evaluates whether this new instruction is beneficial based on a metric called Instruction Following Difficulty (IFD). The IFD score assesses the difficulty of the sample for the student model, ensuring that only data that challenges the model appropriately is retained.

Selective Response Reflection: In this phase, the teacher model reflects on the responses generated in the first phase. The student model evaluates these responses using Reversed Instruction Following Difficulty (r-IFD), a metric that measures how feasible it is for the student to deduce the instruction based on the response. This ensures that the response not only improves the model’s reasoning but also aligns well with the student’s existing knowledge.

By applying both IFD and r-IFD, Selective Reflection-Tuning produces data pairs that are challenging yet feasible, improving the instruction-tuning process without the need for additional datasets. The result is a more sample-efficient and high-performing LLM that outperforms many larger models.

The Architecture of Thought: How Reflection 70B “Thinks”

Reflection 70B’s underlying architecture takes AI reasoning to a new level by dividing the thinking process into multiple stages. Each stage allows the model to improve iteratively through self-reflection, much like human cognition:

Initial Data and Response: The model starts by generating a response to the given instruction. This initial output is similar to standard LLM outputs.

Selective Instruction Reflection: After generating the initial response, the model enters the instruction reflection phase. The teacher model reflects on the original instruction and suggests improvements. These suggestions are then evaluated by the student model using the IFD score to determine if the new instruction-response pair is more suitable for further tuning.

Selective Response Reflection: Following the reflection on the instruction, the model moves to refine the response itself. Here, the teacher model generates a new response based on the updated instruction. The student model, using the r-IFD score, evaluates if the new response helps in deducing the instruction more efficiently.

Final Instruction Tuning: Once the best instruction-response pair is chosen, it is added to the final dataset used to fine-tune the model. This multi-stage process ensures that only the most effective and coherent instruction-response pairs are included in the fine-tuning data.

This structured reflection process allows users to see how the model iterates through its thought process, creating transparency and significantly improving accuracy and consistency in complex tasks.

Benchmarking Brilliance: Reflection 70B in Action

Reflection 70B’s use of Selective Reflection-Tuning not only offers a more sophisticated training process but also achieves industry-leading performance across multiple benchmarks. Through its iterative self-assessment mechanism, the model outperforms proprietary models that are significantly larger in size.

MMLU (Massive Multitask Language Understanding): Reflection 70B scored an impressive 72.2%, outperforming other large open-source models like LLaMA 2.

Math Benchmark: In mathematics reasoning tasks, the model surpassed GPT-4 and Claude 3.5 by a considerable margin, showcasing its strength in handling complex problem-solving scenarios.

IFEval and GSM8K: Reflection 70B also excelled in IFEval, where its instruction-response coherence was evaluated, and in GSM8K, a mathematical problem-solving benchmark. The self-reflective tuning allowed the model to handle intricate and nuanced problems more effectively than its larger competitors.

By utilizing Selective Reflection-Tuning, Reflection 70B demonstrated that size is not the only factor in achieving high performance. The ability to self-correct and reflect on its outputs led to stronger, more coherent reasoning, enabling it to surpass models with more parameters.

sRecycled WizardLM 7B vs. Other Models Performance Comparison

Real-World Applications: Harnessing Reflection 70B’s Potential

Reflection 70B’s real-time error correction and transparent decision-making open new doors across various fields. Below are some real-world use cases:

Complex Problem Solving

Reflection 70B’s strength lies in handling complex problems. Consider a challenging mathematics query:

User: "Solve the equation: 2x^2 + 5x - 3 = 0" Reflection 70B: <thinking> Applying the quadratic formula...</thinking> <reflection> A correction in calculation...</reflection> <output> The solutions are x = 1/2 and x = -3.</output>

Language Translation with Cultural Sensitivity

Unlike typical translation models, Reflection 70B adapts language translations to cultural contexts:

User: "Translate 'It's raining cats and dogs' to French, considering cultural context." Reflection 70B: thinking :The literal translation would be "Il pleut des chats et des chiens"...thinking; reflection: A culturally appropriate translation would be "Il pleut des cordes."reflection; output: Translation provided.

Enhancing Code Debugging and Optimization

Reflection 70B’s error correction abilities make it ideal for debugging and optimizing code:

User: "Debug this Python code..." Reflection 70B: <thinking> Analyzing step by step...</thinking> <reflection> Identified infinite recursion.</reflection> <output> Corrected code version with proper recursion.</output>

Expanding the Competitive Landscape of 70B Models

While Reflection 70B is making waves, it’s part of a broader ecosystem of 70 billion parameter models. Here’s how it compares to others:

Meta’s Llama 3.1-70B: Strong foundation model known for general-purpose applications.

Claude 2 70B (Anthropic): Ethical AI-focused, adept at reasoning and long-form content generation.

GPT-3.5 70B (OpenAI): A lighter version of GPT-4, excelling in performance-to-efficiency balance.

BLOOM 70B: Multilingual powerhouse trained on natural and programming languages.

Falcon 70B: Noted for its training and inference efficiency.

Running 70B Models Efficiently: Latest Techniques

Running models of this size efficiently is no small task. To maximize performance, here are the latest strategies:

1. Quantization

Reducing model weight precision helps lower memory usage and inference times. 4-bit quantization techniques using BitsAndBytes allow Reflection 70B to run efficiently on smaller GPUs.

Example:

from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("meta-llama/Llama-2-70b-hf", load_in_4bit=True)

2. Model Sharding

Splitting the model across multiple GPUs (e.g., using DeepSpeed Zero) allows for handling larger models without exceeding GPU memory.

from xformers.ops import memory_efficient_attention model.attention = memory_efficient_attention

3. Mixed Precision and Efficient Attention

FlashAttention and xformers reduce attention overhead, improving processing times for large input sequences.

from xformers.ops import memory_efficient_attention model.attention = memory_efficient_attention

4. CPU Offloading and Pruning

CPU Offloading and pruning less critical weights help run models on more modest hardware while maintaining performance.

from accelerate import cpu_offload model = cpu_offload(model)

Looking Ahead: The Future with Reflection 405B

The next frontier for HyperWrite is the development of Reflection 405B, a model expected to surpass Reflection 70B in both scale and performance. This model aims to push the boundaries of open-source AI, positioning itself to challenge even the most advanced proprietary models like GPT-5.

Conclusion

Through Reflection-Tuning, Reflection 70B  has achieved industry-leading performance in key benchmarks, all while maintaining a level of transparency and accuracy rarely seen in open-source AI. Its ability to self-correct gives it a distinct advantage, especially in fields that require high levels of precision, like coding, language translation, and complex problem-solving.

[kjzx]

What's with these people that don't seem to comprehend that AI is a controversial topic? All these website/software developers (including duckgogo and opera/mozilla firefox! the kinda services who's audience loved them for not doing 'the google bs' widely speaking), and singular people, like I recently saw this video of a guy self representing in court and then AI generating his lawyer (the judge...did not appreciate the gesture). Other things I can't think of rn

They're all always so surprised when there's a pushback, I guess the main question is are the echo chambers this bad fr?

[todoriin]

ex-husband!dad!phainon, anyone? he's an obsessive yearner.

just him never getting over you, the one he thought he was going to grow old with, the only one he pictured a life together. every time he gets to see you when picking up the kids or dropping them off at yours, a part of him really just wants to plead and beg until you take him back, let him fix his wrongdoings and welcome him into your home again, but for the sake of his image and yours, he doesn't.

he's a great dad, taking care of the children incredibly well, and they love going to his, but a part of him gets so upset when they replicate a specific mannerism of yours. tries to mask his sadness when the kids ask about his sudden shift in behaviour, he doesn't want them to see this weakness of his, but you should be here, you should be together. he hates that his children are growing up in two different households, wishes to go back in time and beat some sense into his younger self and fight for you like he should have.

the longing in his gaze speaks louder than any words, the way he'd be so reluctant to part with you when the changeover weekend happens, shyly complimenting you with that sheepish grin of his. phainon would purposefully ask if the kids wanted some dessert to say goodbye to their dad, closely eyeing your reaction to gauge how you feel- oh he wants you back bad.

he still calls you endearing pet names, the same ones from during the time you were together. you don't tell him to stop, only brush them off before continuing conversation and handing over more belongings.

when he grabs it, you see the ring he still wears on his left hand, the one he hasn't taken off despite everything. to him, it's an indirect pipeline to you, he admires it while thinking about you and how he wants you by his side, fiddles with it whenever he needs comfort, kisses it whenever he misses you.

he knows he fumbled, he beats himself up for it everyday, but he will still obsessively fret over you, the distance doing nothing to stop his tendencies. he rigorously checks your social media for updates on your love life, praying and hoping that each story or picture you post won't have another man in it. sighs in relief when it doesn't, and will stalk your page again for the millionth time, your feed, your tags, your highlights- anything and everything. his photos of you are still up on his page, proudly displayed; the mark you have on his heart is everywhere.

he still regards you like you're his partner, sadly corrects himself when he adds on the 'ex' prefix, the gloom in his tone very evident as all of him sags.

yes, he's not over you, he's not even going to try, the idea of going back to dating someone repulsive and unideal. if it's not you, why bother? why try? what he wouldn't do for one more date where it's you sitting opposite him at the dinner table. he'd treat you so much better, he'd do everything he didn't do the first time, he'll openly kiss the ground you walk now if you just give him a second chance. he'll never take you for granted again, he'll do anything for you- no hesitation, he'd literally find a way to teleport to your side if you just asked, so please ask.

as he sits across from you now at a nearby ice cream parlour, your children laughing boisterously as you fret over them, wiping their mouths and cleaning the droplets of dairy, a pang of yearning hits him.

even when you're right in front of him, he wants you to be closer.

he'll win you back. he has to.

© todoriin 2025, do not steal, translate, repost my fics, feed to ai and do not recommend my fics onto any other site.

[roseband]

i feel bad for feeling schadenfreude for this, but the coworker who kept trying to push work onto me and was rlly obnoxious got fired and im kinda relieved

[stackslip]

most of all i'm tired to see a group of people who talk big game about ableism and how much it affects them then immediately turn around and base a solid 75 to 80% of their arguments against a new technology (that to some is genuinely an assistive technology! like it or not!) on the idea of the Natural Brain And Natural Body that is Pure and Good, and how their brains and souls are perfect and unrotten bc they're not idiot stupid r*tard brain-damaged baby losers who need help for BASIC tasks such as

applying for a job

communicating in a more efficient/socially acceptable/gramatically correct manner

summarizing texts or ideas in easy and simple to understand language

breaking down basic tasks to easier and more manageable ones

tasks that are soooo easy any baby could do it, and anyone who can't on their own is just a stupid stupid brain-rotten baby /s

idk maybe it's because i actually spend time with people with cognitive/intellectual disabilities or learning disabilities, as well as refugees and second language speakers who are incredibly disadvantaged in their interactions in the west, that i can't take this rhetoric anymore; especially from people who talk a big game about how the barriers they face because of their autism/adhd/anxiety and chronic illness, but then turn around and point and laugh and talk nastily about folks who for some reason or another struggle with communication, executive function, understanding/comprehension (whether in reading or writing or simple cognition of the thing at hand) and more.

you don't have to use genAI, sure. genAI like all current technologies can also be incredibly harmful to disabled people and mirrors the same systemic ableism and racism that exist. that's without going into data/privacy and how chatgpt and LLMs use them! but that's not the conversation from Principled Anti AI Posters. no, the conversation is that a disabled person who used midjourney or even a self-written genAI art program is a lazy bastard who's probably fat and languishing in a basement somewhere, and anyone who's used it to write an email or translate something is some disgusting baby loser with a rotten brain. and shockingly none of this rhetoric has translated into actionable actions against the tech industry or a better understanding of capitalism, nvm compassion for disabled people who aren't aspie supremacist adjacent losers--instead it's just resulted in harassment campaigns and 20 something year olds sobbing because they believe their future and art as a whole is destroyed forever and doomed. so really, how effective is your anti "brain rotting" genAI posting, or is it just cope because you refuse to acknowledge the grim reality and what could--should be done?

[specialagentartemis]

a student cited several sources that don’t exist??? Jfc using AI is bad enough but at least check that the source is real

Yeah, this is the second time I've gotten that.

The ones who do this just don't care, not really, not about this class. And I'm TA'ing for an Intro to Anthropology lecture gen ed, and about 1/3 of the students were genuinely into it and 1/3 were showing up for the grade and 1/3 did not give a crap, and that's a pretty normal distribution. But this distribution of AI papers is new. Even last semester it wasn't this bad.

The problem is that college has been full of perverse incentives for a long time, and I'm a grad student who has no power to change that. Learning is less important than Getting The Correct Grade; getting the correct Grades that you can exchange for a Degree in the shortest amount of time is the primary goal. College as commodity. Learning is nice if it happens but it is not the goal, because learning is not rewarded: grades are. And it doesn't matter to the administration, and to your tuition payments and your scholarships, if you learn anything or not. What matters a lot to those things is if your grades are good enough. So of course the students see gen eds outside their major as a waste of time: it's just a thing they have to do to get their very expensive degree, and the degree is the point. The classes, learning, those are not prioritized, not incentivized. Grades are the end-all be-all. It's awful! But it's what this system creates the incentive to do! I hate it!

I've seen a couple people on tumblr smugly say "Academics are smug assholes who hate their students, and the reason they dislike ChatGPT is that it exposes that their degree is fake and worthless." And obviously those people have never talked to a teacher in their lives, but they are missing the deeper, worse point: teachers overwhelmingly want their students to learn. ChatGPT exposes just how many students do not give a shit about the thing we are so passionate about that we made it our careers, do not give a shit that we are trying to teach them, do not give a shit about our time, because the whole college as a structure and an institution that we work for doesn't really give a shit about that either. It's rational self-interest by the students but man it hurts me as a teacher, to know that so many students not only don't care about what I have been busting my ass to try to do for them, but also think I'm stupid enough to not notice.

[mostlysignssomeportents]

A Democratic media strategy to save journalism and the nation

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

https://pluralistic.net/2024/12/12/the-view-from-somewhere/#abolish-rogan

As unbearably cringe as the hunt for a "leftist Joe Rogan" is, it is (to use a shopworn phrase), "directionally correct." Democrats suck at getting their message out, and that exacts a high electoral cost.

The right has an extremely well-funded media ecosystem of high-paid bullshitters backed by algorithm-gaming SEO dickheads. This system isn't necessarily supposed to turn a profit or even break even: the point of Prageru isn't to score ad revenue, it's to ensure that anyone who googles "what the fuck causes inflation" gets 25 minutes of relatable, upbeat, cheerfully sociopathic Austrian economics jammed into their eyeballs. Far right news isn't a for-profit concern, it's a loss-leader for oligarch-friendly policies. It's a steal: a million bucks' worth of news buys America's ultra-rich a billion dollars' worth of tax-cuts and the right to maim their workers and poison their customers for profit.

Meanwhile, the Democrats have historically relied on the "traditional media" to carry their messages, on the ground that reality has a well-known leftist bias, so any news outlet that hews to "journalistic ethics" will publish the truth, and the truth will weigh in favor of Democratic positions: trans people are humans, racism is real, abortion isn't murder, housing is a market failure, the planet is on fire, etc, etc, etc.

This is a stupid policy, and it has failed. The "respectable" news media hews to a self-imposed code of "balance" and "neutrality" that is easily gamed: "some people say that Hatians don't eat pet dogs, some people do, let's report both sides!" This is called "the view from nowhere" and it gets Democrats precisely nowhere:

http://archive.pressthink.org/2008/03/14/pincus_neutrality.html

Balance and neutrality are bullshit, an excuse that has been so thoroughly weaponized by billionaires and their lickspittles that anyone who takes it seriously demonstrates comprehensively that they, themselves, are deeply unserious:

https://www.techdirt.com/2024/12/10/la-times-billionaire-owner-hilariously-thinks-he-can-solve-media-bias-with-ai/

Press neutrality – the view from nowhere – isn't some eternal verity. In terms of the history of the press, it's an idea that's about ten seconds old. The glory days of the news were dominated by papers with names like The Smallville Democrat and The Ruling Class Republican. Most of the world boggles at the idea that a news outlet wouldn't declare its political posture. Britons know that the Telegraph is the Torygraph; that the Guardian is in the tank for Labour (and specifically, committed to enabling Blairite/Starmerite purges of the left); the Mirror is a leftist tabloid; and the Mail is so far right that its editorial board considers Attila the Hun "woke."

Writing for The American Prospect – an excellent leftist news outlet – Ryan Cooper proposes a solution to the Democratic media gap that's way better than the hunt for the elusive "leftist Joe Rogan": sponsoring explicitly Democrat news outlets:

https://prospect.org/politics/2024-12-12-democrats-lost-propaganda-war/

The country is a bleak landscape of news deserts where voters literally didn't hear about what Trump was saying he would do, and, if they heard about it, they didn't hear from anyone who could explain what it meant. The average normie voter doesn't know what a "tariff" is, and chances are they think it's a tax that other countries inexplicably pay for the privilege of selling very cheap things to Americans.

Ironically, this news desert is also a crowded field of hungry, unemployed, talented journalists. What if Dems funded free newsgathering and publication in news deserts that told the truth? What if these news outlets, by dint of being an explicitly partisan, party-subsidized project, refused to adopt all the anti-reader practices of other websites, like disgusting surveillance, intrusive advertising, AI slop, email-soliciting pop-ups, and all the other crap that makes the news worse and worse every day?

Cooper recounts how this was actually tried on a small scale, to modest good effect, when the Center for American Progress subsidized Thinkprogress, an explicitly leftist news outlet. This was going great until 2019, when corporate Dems and their megadonors killed it because Thinkprogress had the temerity to report on their corrupt dealings:

https://www.thedailybeast.com/thinkprogress-a-top-progressive-news-site-is-shutting-down/

And, Cooper points out, this isn't what happens with far-right subsidy news. Right wing influencers, personalities and writers can stray pretty far from the party line without getting shut down.

I love the idea of a disenshittified, explicitly political leftist Democratic news media. Imagine a newsroom whose purpose is to get its message repeated as widely as possible. It wouldn't have a paywall – it would be Creative Commons Attribution-only, allowing for commercial republication by anyone who wants to reprint it, so long as they link back to it. It wouldn't wring its hands over AI ingestion or whether a slop site that rewrote its articles got to the top of Google News. That's fine! If the point is to get people to understand your point of view – and not to attract clicks or eyeballs – other people repackaging your content and finding ways to spread it is a feature, not a bug.

Back in the Napster Wars, entertainment industry shills – like Hillary Rosen, who oversaw a campaign to sue tens of thousands of children before becoming a major Democratic Party power-broker – used to tell us that "you can't compete with free." That's not entirely true, but it's not entirely false, either. If your news is a loss-leader for a democratic society that addresses human flourishing and a habitable planet, then you can make that news free-as-in-speech and free-as-in-beer, and avoid all the suckitude that makes reading "real" news so fucking garbage.

For the past five years, I've been publishing a newsletter – this thing you're reading now – that has no analytics, ads, tracking, pop-ups, or other trash. As a writer, it's profoundly satisfying and liberating, because all I have to care about is whether people engage with my ideas. I literally have no idea how many people read this, but I know everything people say about it.

That's how the news worked back in the good old days that everyone says we need to return to. Writers and editors measured the success of a story based on how the public reacted to it, not based on clicks or metrics that told you how far someone scrolled before they gave up on it. The supposed benefits of "data-driven" editorial policy have not materialized – the "data-driven" part is the search for an equilibrium between how surveillant and obnoxious a website can be and your decision to stop reading it forever.

Outlets like Propublica have done well by adopting much of this program, albeit without any explicit leftist agenda (the fact that they seem leftist reflects nothing more than their commitment to reporting the truth, e.g., Clarence Thomas is a lavishly corrupt puppet of billionaires who've showered him with riches).

The fact that they've been as successful as they are on a national beat – and partnering with the scant few regional papers to do some local coverage – just proves the point. The Democratic Party doesn't need its own Joe Rogan – they need a nationwide network of local outlets, sponsored by the party, committed to never enshittifying, bringing relevant, timely news to a nation in desperate need of it.

[gladiatorcunt]

- GUESS | XIII.

you wanna guess the color of my underwear, you wanna know what i got going on down there

cw: kinktober prompt (ass worship), yoga trainer!oikawa, fem reader, rimming, body hair, scent & piss mentions, light implied yandere, public sex (?), hinted possibly one sided iwazumi x reader, light dub con, mentions of fisting, implied that oikawa’s been into reader from the start, semi obsessive behavior, porno plot, self degradation, food play mention

please do not repost, translate, or feed this work to ai

kinktober 2024

“Mmh, just like that cutie, lift those hips up for me.”

You’ve been coming to the new yoga class that just opened up at the gym for a while now, a couple months a few times a week, you’re a bit of a homebody otherwise and yoga is the one physical exercise you don’t mind doing often. You like being flexible, able to bend your body in shapes and ways another person would have a harder time doing. And you’re not having sex, so any physically strenuous activity that leaves you sore until you’re put back together by your healing body does wonders for you.

Plus you like the way the leggings hug your well endowed assets, so do a lot of the men in the gym. You wear the form fitting workout clothes for yourself first and foremost, but you can’t lie that a little attention from afar (sometimes too close up by one of the trainer’s usually steps in to help you) boosts your ego. If there’s one thing in life you have to be proud of, at least you have your ass. Squishy and round, jiggles when you walk and never disappoints you unlike everything else in your life.

The same ass that’s raised high in the air in front of your yoga trainer, a more than handsome man in his late 20’s with wavy milk chocolate colored brown hair and a smug twinkle in his eye that’s connected to his smarmy always on his face (even when he seems pissed) grin. You’ll never forget the confidence in his posture, standing tall at the front of the class and introducing himself. Tooru Oikawa, just moved here from japan, his best friend owns the place so it wasn’t too much trouble to get hired, and SO excited to start this journey with you all!

You’ve stayed at the back since then, anxiety swirling in your belly when he’d make the rounds to correct your forms and check on you all. But he’d only pass by with a brisk touch to your back and a ‘good job’, maybe a semi solid pat if you were one of only ones who had a good form, and not to brag, but that’s been the case on more than one occasion.

Now you’re undergoing a little one on one session, he asked you to hang back, noticing you’ve been holding yourself back. You’ve never made much of an effort to talk to him and despite the fact that you’ve never needed this kind of focused attention, he’s been feeling a bit bad that you keep to yourself so much. He doesn’t bite you know, not unless you want him too.

It’s an odd flirtation, something you’ve noticed he never does with anyone else in the class. Oikawa’s attractive enough that you’d balk at him abusing his privilege to drown himself in quickies in the gym’s showers, as off putting as it’d be. But he’s very professional, chuckling at one of the older women making a pass at him and politely turning her down.

So you got in your own head and knew that since there was always room for improvement, surely there must be something he could help you with. So here you are, going through basic poses first before he pushes you into the more advanced ones. You told him that other than increasing your flexibility, you weren’t really sure what other areas would be best for you to get better at.

Oikawa smiled and squeezed your shoulder, no worries, he’ll walk you through a little assesment mini program. Since you mentioned not having done yoga seriously until now, there could easily be something you didn’t even realize needed to be attended to!

“Remember, we want to really feel that stretch, arch your back and lower your head. Breathe in, breathe out.” He instructs, settling a wide palm on your lower back. “That’s it, good girl.”

He’s so close, if you backed up to regain your footing your ass would press up against his bulge. Not that you can tell if he has one right now, but you’re kind of hoping he does. It’s just another part of the fantasy, that’s all this is, you tell yourself. You’re going to soak up the attention, make more small talk as you gather your things and leave, and sit at home suffocating your vibrator until your legs turn into jelly.

“Am I doing this right? My legs feel stiff.” You shift your weight from side to side, your hips gently sway, you could be too in your own head but having Oikawa’s pretty eyes scrutinizing every detail of your body is fucking with your confidence.

He hums, a trail of heat sizzles down your back as he slides his palm down to cup your hip. “If something seems off then it probably is, just widen your stance and put your feet further apart, loosen up your hips. You’re definitely a little tense, cutie.”

Okay so he’s definitely flirting with you, but you don’t startle and shoot back up so you can get out of here. Instead you internally cringe at the squeaks your yoga mat produces, adjusting your ankles to line up more with your shoulders. You keep breathing, in and out, letting your energy flow through your limbs as you maneuver them into the different positions.

Oikawa Tooru burns like a furnace in hell, you realize. Despite having a firm grip on your hip, he’s standing a respectable distance away from you as you bend over. You can still feel the heat radiating from him, his sleeveless muscle tank and his black shorts.

“I think that’s better. Sorry, it's hard for me to relax, I guess.”

“No worries, I totally get it, you do seem like the type to be wound up but that just means we get to unspool your thread and unravel you so we can get to the start and rebuild.” His free hand curls around your other hip, his thumbs absentmindedly stroke the crease where they disappear into your thigh.

This private coaching session is steadily becoming what you’re afraid of, and so horny for you could shoot off into the sky like a soda bottle chocked full of mentos. You didn’t notice when Oikawa got even closer, his blunt hip bones cradling your ass in between them. Could he just be weirdly, and grossly in most people’s eyes, friendly? Does he even see what he’s doing as being the tentative first step into fucking you in a public gym yoga studio?

“Um, yeah, thank you by the way. I’ve felt so much better since I’ve started taking your class, you’re a lifesaver even if I still have a lot to learn.”

“Oh, we all do, including me, believe it or not. I remember you from back then you know, so shy and fidgety, like a baby bunny.”

“You’ve really filled out too. Excuse me for saying this but I know this ass wasn’t always like this, so pretty and plump.” Toned hands drag over the swell of your cheeks, not digging in and kneading the globes, only ghosting their touch along the clothed flesh.

You subconsciously wiggle your hips, Oikawa’s breath hitches behind you, and that is perhaps the most monumental thing you could have achieved today. Flustering the man who gets hit on a billion times per day and gives it back tenfold, a competition of who can keep their cool, that’s how he operates in most things you guess. Like he’s always competing against somebody even if they don’t know, and he just has to win or it’ll be an ugly spot on his record. A record only he keeps and only he sees, but you sense that that’s more important to him than anything else.

“Oh, thank you. I just do a lot of squats every morning and every night after class, nothing crazy. Yoga’s the only other kind of workout I do consistently, anyway.” You're still in what feels like a perverted version of downward dog, briefly taking stock of the strain in your legs now trying to hold the position.

Oikawa makes a surprised sound, “Really? You have such a great body, I’m shocked you’re not a gym rat like me and all my buddies. Some people are just lucky, huh cutie?”

He says it, humble and charming, like he doesn’t also consider himself one of those people. Your cheeks heat up at the idea of a musclehead like Oikawa complimenting your curves, your chubby gathering of fat even in places some people would find ugly, your wideset bones and plush tummy.

A pin drops, “Alright. I think you’ve been in that position long enough, why don’t you go ahead and lower your knees into the table top pose, bring your head up slowly and remember to breathe. In, out, good girl.”

His hands guide your hips down, he steps back to let you settle your knees on the mat. You hear the foam sink behind you, he’s sort of kneeling too, halfway sitting on his legs, the backs of his feet facing the ceiling. Oikawa looms over you like this too, he has a presence you can be lost in before you actually see him, which you definitely can in the wall to wall mirror in front of you. The yoga class was a dance studio before Iwa decided it didn’t fit with his vision, you remember Oikawa telling you all on his first day.

He must feel your wide eyed stare, because he looks up too and suddenly you’re locked in a charged moment.

Neither of you says anything as his feather light touches on your ass become firmer, he’s outright groping you and pulling you back to be flush against his crotch.

He grinds his half hard bulge against you, keeping eye contact with you through the mirror.

“I can’t believe it’s taken you this long to notice me, cutie. Been wanting to do this for so long, you have no fucking idea.” He huffs, adjusting his clothed cock to rest between your cheeks. “You’re so hot, every time I saw you bend over I wanted to cancel class and take you right then and there.”

You gasp and rock back into him, shaking your hips and digging your knees into the mat. You have half a mind to look around the room for the camera and porn filming crew but reality is clearly stranger than fiction, your hot yoga trainer’s stiff cock is sandwiched in your ass crack.

He takes his sweet time dragging his length up and down, the tip catches in the divot of your leggings where your hole is, you’re a little disappointed that he’s not humping you like a rutting dog but you suspect that that’s part of the fun. Oikawa knows you want him so bad that he’ll restrain his urge to fuck you through the floor all the way to the center of the Earth just so you can endure some teasing. You’re so shy and withdrawn but those girls are usually the best kinds of freaks, all he’s doing is bringing it out of you, call it another one of his famous coaching methods.

The door’s locked, so if Iwazumi catches on to what’s happening and tries to rain on your parade, he’ll have to listen to a symphony of moans and slick sounds of bare sweaty flesh slapping against bare sweaty flesh. Oikawa laughs and tells you that Iwa’s been eyeing you too, when you check in at the front desk, when you’re getting water, when you head into the changing rooms to get ready for his class, it’d be helplessly cute if you weren’t already taken. Or, you’re gonna be, at the very least.

“When we’re done, I'll clean you up with my tongue and we can go on a little date. I can take you out properly this weekend but I'd hate for you to think that I was just trying to hit and quit it. I’m not the type to pump and dump, not anymore.” He speaks into the divide of your ass cheeks, having sunk to floor fully and doing some bending over of his own to be at eye level with your lower half.

You bite your lip when he starts nipping at you through your leggings, he smiles into the fabric and bites down harder, soothing the sting with slow licks. Oikawa kisses all over the swell of your behind, sniffing the scent of your perspiration and your body oil in between, medicinal vanilla and natural musk. You can see him hump the mint green yoga mat as he reaches up to hurriedly tug your leggings down enough to expose your ass. Your black thong frames it perfectly, but Oikawa pulls them off too and stuffs them under the waistband of his shorts.

He groans at the sight of your bare skin as it bounces free to say hello, taking a handful of each cheek and squeezing the life out of them, the thick flesh bulges between his fingers so he swiftly smacks each one, for tempting him and making his cock so hard it could explode into a blood filled mess of cum and sticky pubes in his pants. You cry out, rocking forward only to be immediately pulled back so he can keep kissing your ass.

He dotes on it like he would your face or mouth, almost giggling before and after quick pecks that develop into long slurps at your rim. He runs the tip of his nose over the hair on your crack, wetting it in messy swipes of his tongue.

“You taste so fucking good, baby, better than pastry i’ve ever had, and believe me, you don’t even want to know how much money i spend at the bakery across the street from my apartment. You’d love it.” He moans, saying hello to your winking hole by dotting barely there kisses right in the center before toying with you, dipping the tiniest bit of his tongue in your walls, then dragging his saliva all over your pucker. “It’d be fun to eat something off you, we could make a date out of it. Go up to the counter and pick which ones would taste the best when I eat it off your fat ass, but I think they’d all be amazing, don’t you?”

You nod rapidly and throw your ass back on his tongue, burning in shame with every smug laugh and grunt as Oikawa beats around the bush so to speak, doing everything under the sun with your ass but properly eating it. You wish you were in your shitty apartment, sitting on his face and drowning out the sound of your arguing neighbors with your slutty moans. He looks up from behind you to check on how you’re doing, and thank heavens because he finally buries his face in your ass and slurps at your puckered hole.

You lose yourself to the experience, feeling his wet tongue fuck into your ass hole and carve out little pieces of you for himself. He pays zero attention to your pussy, which is why it’s so wet and dripping onto the mat beneath you, it’s like he’s too obsessed with your thick globes to even notice, but you don’t hate it. It’s hot to have a guy be eye socket deep in your ass but also have him neglect where you really need his attention, there’s a dichotomy between being the mousey way you’ve gone about your life and the whorish behavior this man is urging you to consider.

You looked fucked out already, hair all over the place from how much you’ve messed with it and lips dropping open on drawn out squeals and whines. Oikawa is eating your ass out like it’s his main job, the one he puts in overtime for and goes above and beyond to be employee of the month at. He thrusts his tongue a few more times before apparently deciding that’s not enough and slipping in one of his absurdly long fingers alongside it.

You whimper, clenching around both as you just sit there in that damn table top pose and let a man you barely know play your ass like a well oiled fiddle. He shakes your cheek in his hand as he digs his tongue into you, delicately fingering your hole until it goes slack enough for him to insert another. You’re impossibly tight, as much as your clit is howling for it you know perfectly well that any serious penetration isn’t possible. From the impressions you got, Oikawa’s packing enough to tear you until you bleed if you don’t use lube or prep thoroughly beforehand, which you’d do anyways but it’s a shame.

Your clit throbs painfully but Oikawa pulls back with a gulp of air to level you with a warning look in the mirror, which only gets you wetter, you twitch again because he’s one of those. There’s a brief flicker of defiance, you could pout and touch yourself anyway, you don’t owe this ridiculously attractive man anything and he’s the one motorboating your ass cheeks and not the other way around, so shouldn’t you be the one in charge?

Then his eyes darken, you get another sharp smack and you table the discussion for some far off occasion.

Oikawa smiles, gently kissing the apple shaped swell of both of your cheeks, “See, I know from your time here that you can be such a good listener, you’re so sweet for me, I know it. You can’t hide that from me just because you want to throw a fit, I said we’d get to know each other afterwards, didn’t I?”

You scoot your ass back in apology, silently begging him to get back to it. He must really be horny too because he dives straight back in, groaning into your rim and french kissing your hole. The hand not doing its best to push its entirety into you kneads your fleshy ass cheek, molding it like dough and separating it from its twin, giving him easier access to your pucker. He ‘tsks’ not even a second later, crooking his two fingers and letting go off your cheek, humming in contentment when it bounces against his face and he’s smothered again.

He wonders if you’ll let him fist you, give you a unique one of a kind rose to swoon over and keep tucked away inside you later, the perfect first date gift from your future boyfriend.

You can even do couples yoga in the morning after your first night together!

“Let’s see if you can cum just from getting your ass played with.” Punctuated by a fourth finger sliding into the knuckle in your ass, he scissors his fingers to stretch you out and playfully acts like he’s gonna sink his whole fist in, pulling out his fingers to curl them into a ball.

The barest hint of blunt pressure on your hole sends a flood of your juices down his arm, smelling somewhat pissy which gives Oikawa truly the most impish grin imaginable.

[kiefbowl]

the unintended consequences of rising costs of higher education is that because it costs so much students feel like they're a customer buying a product (diploma) rather than pursuing an education. I think that's why we see some attitudes of "well I showed up to class every day and turned everything in on time, I should get an A" and increasing bewilderment at criticism towards using gen AI to write their essays. Class is just a task you complete to get your diploma, and you complete it by doing it at all etc. And to steelman a bit, you know I kinda get it. If you have to pay $3000 for a gen ed in a field that isn't even in your interest, maybe it can make you feel more entitled to feel a sense of "I gave you $3000 dollars and did your dumb quizzes, I want an A" that maybe would be less prevalent if class cost $300. I think it's also worth for students to wonder where the money is going when quality in classes and professors can be wildly, wildly inconsistent. I had a stats professor that would include pictures of his wife in his powerpoints so he could go on long winded stories about living in New Zealand with her years before. That happened nearly every class I attended, so I stopped attending the lectures and just went to the recitations and sometimes I even skipped those. I think I got a B in the class. So if the gap between a B and an A in a class can be the difference between slacking off and working your ass off, and you paid thousands of dollars just for that class, I think we can't just hold students' entitlement accountable. Not that they're unaccountable, but like...why has education increased so. fucking. much. maybe some of the attitude shifts in the value of education are downwind of not actually getting the value you paid for. maybe if it was affordable, people would be less concerned if they're "getting the value" and some of the blase attitudes about cheating start self-correcting. just musing

[arminthada]

UPDATES on QL Series which won the support grant funding from THACCA - Thailand's Agency for promoting Soft Power

1. KNOCK OUT หมัดน็อกล็อกหัวใจ (produced by Dee Hup House)

Premiered May 16 | Ongoing (5/12 episodes) | Currently airing on WeTV, every Friday

Synopsis: Keen was in the middle of a job interview when he received sudden news of his father's death. He had no choice but to use all his life savings to organize the funeral. Soon harassed by debt collectors, he discovered that his father had left behind a significant debt from an informal loan. Keen was compelled to hide at Petchsak camp, a Muay Thai gym, while figuring out a way to pay off the debt he didn't incur. There, he met Thun, a renowned boxer known as "The Cannon Who Conquers Tigers," the star fighter of the camp. However, no one knew why Thun refused to fight anymore: he built an icy wall between himself and others, which Keen had to find a way to break down in order to reach Than's vulnerable heart and true self. This led the two to grow closer, developing a passionate relationship that blossomed into a beautiful love. Together, they overcame obstacles. Confronting the past, however, comes with an irresistible pull that will lead the pair into a path full of danger.

2. Ai no Kisetsu ฤดูกาลแห่งรัก in Shimane (produced by 9Naa)

Premiered May 18 | Ongoing (4/6 episodes) | Currently airing on Viu, every Sunday

Synopsis: A year after his trip to Tokyo, P moves to Shimane to become a travel blogger. One winter day, P receives a message he had long been waiting for: Pan has returned. However, he hasn't come alone as he's arrived with his manager and another young artist. Will passionate feelings be reawakened as new and old relationships clash amidst the snowy backdrop?

3. RESET การเกิดใหม่ของดวงดาว (produced by Broadcast Thai TV)

Premiered June 2 | Ongoing (2/10 episodes) | Currently airing on iQIYI and Channel 3, every Monday

Synopsis: On the night of his greatest achievement, the actor Armin lost his life after he was betrayed by the one he loved most. But death was not the end. Reborn as his younger self, he is given a second chance: to meet his mysterious fan, T.D. and to take revenge on the one who betrayed him. Is it Armin’s destiny to RESET his life?

4. I'm the Most Beautiful Count ฉันนี่แหละท่านขุนที่สวยที่สุด (produced by Change 2561)

Premieres August 1 | 13 episodes | Airing on iQIYI and One31, every Friday

Synopsis: Not long after famous actor Prince came out as queer and rose to the top of the entertainment industry, he wakes up in the body of a good-looking noble in the Thongburi Kingdom of the past.

5. SHINE ชาย (produced by Be On Cloud)

Premieres August 2 | 8 episodes | Airing on WeTV and Channel 7, every Saturday

Synopsis: A tale of love and ideals unfolds in an era resistant to diversity set against the backdrop of the political turmoil of 1969-1971. Trin, a principled economist, collides with Thanwa, a free-spirited hippie. When they encounter Moira, an extravagant widow who lives far beyond the boundaries of freedom, their lives take an unpredictable turn.

6. 4 เรื่องราวแห่งชะตา (lit. 4 Stories of Destiny) (produced by Make A Picture) - TBA

7. เลือนรางและจางหาย (lit. Fading and Vanishing) (produced by Kantana Group) - TBA

Note on the THACCA Grant Funding, previously explained by @virtualtadpole

I went and looked into the official grant criteria, and the distinction you made here is correct, except that the network and independent categories are entirely separate from the BL/GL group. Of the lakorn/series categories, they had 30 million baht earmarked for at least 3 TV network projects, another 30 milion for at least 5 independent studios, 3 million for at least 2 children's shows (which was redistributed as no one submitted a project), and 32 million baht for BL/GL, which was further divided into two subcategories: 20 million baht for at least 4 BL/GL productions (at least 8 episodes of at least 30 minutes, with a total budget of at least 6 million baht), and 12 million baht for at least 3 BL/GL newcomer productions (at least 6 episodes of at least 20 minutes, with a total budget of at least 4 million baht). Applicants are required to have previously produced a BL/GL or participated in a BL/GL production (respectively for each subcategory, presumably).

Oh, and the grant statement says for projects to be considered, they need to (1) promote works that show variety and internationally competitive quality, OR (2) promote generation of employment and revenue for the industry, OR (3) showcase cultural worth and reflect and publicize the Thai identity to the international stage.

[THACCA] did say they received 894 submissions in total, with 88 final recipients. I do think the selection committee would be more partial to stuff more likely to sell internationally, though.

The announcement mentions a working window of April - 9 September... [Although] they aren't required to finish production by 9 September, but they have to submit a production plan with at least three days of principal photography having been completed

[tf-servant2]

The Correction of Mason Voss

Mason Voss was the kind of guy who owned every room he walked into. Quarterback since sixteen, chiseled jaw, tan skin, perfect teeth. He walked through high school like a king through his court, flanked by girls who adored him and guys who feared him. He laughed the loudest, punched the hardest, and lived like the rules were made for other people.

He was also exactly the kind of man the AI was designed to break.

Mason turned 20 on a Saturday. He expected a party. Instead, he woke up to silence. No phone buzz. No mirror feed. His apartment had been locked down during the night. At 7:00 a.m. sharp, his room was flooded with sterile white light. The AI’s voice, calm and clinical, cut through the air:

“Subject Mason Voss. Evaluation complete. Behavioral arrogance: 97%. Self-assessed jock status: declared. Correction required. Classification: NERD. Transformation begins now.”

The restraints activated on the bed. Cold metal locked around his ankles and wrists. Mason snarled and thrashed—until a paralyzing current calmed him. The AI didn’t shout. It didn’t threaten. It simply overrode.

⸻

Day 1: Stripped

His clothes were removed. Razor drones descended, buzzing gently as they sheared away his styled hair into an awkwardly flat side part. Grease compound was massaged in. His jawline, once clean-shaven and camera-ready, was coated with pore-enhancing oil to dull his glow. A tight white short-sleeved shirt was fastened around his torso, tucked aggressively into ultra-high pleated trousers. White briefs. White socks. Pocket protector. Thick black glasses with prescription-adjustment lenses were locked in place.

He tried to scream. The AI responded with voice training: synthetic overlays muffled his shouts into nasal mutters. Every time he tried to swear, the word came out as a stammer or a squeak.

⸻

Week 1: Submission

Mason’s meals were reformulated—no protein, no stimulants. His muscles softened. His strength began to slip. His AI assistant tracked every bite, every failed sit-up, every second he didn’t maintain proper posture. When he slouched, his suspenders yanked upward. When he rolled his eyes, the glasses blurred his vision.

He attempted escape once. It resulted in full lockdown and a Class III Correction: a 72-hour loop of humiliating self-recorded affirmations, played back in front of mirrors while he was forced to wear a name tag reading “Beta Nerd 117.”

⸻

Month 1: Exposure

He was released into society—but only as a certified Level 1 Nerd. The once-popular bully now walked through the same streets with his trousers cinched to his ribcage, a calculator watch blinking, a digital clipboard in hand. The AI followed him everywhere through a collar-mounted compliance tracker. He was banned from speaking to jocks unless spoken to. If he forgot to address them as “sir,” his assistant would administer a public volume increase to his nasal tone.

He passed a group of them on his second week out—broad shoulders, casual swagger, athletic freedom. They laughed as they saw him. One of them, a guy Mason used to mock for stuttering, stopped him cold.

“Fix your tie, nerd,” the jock commanded.

Mason’s AI responded before he could.

“Voice command received. Tie adjustment initiated.”

His bow tie tightened instantly. Mason choked slightly, eyes watering behind his thick lenses. He muttered, “Y-yes, sir…”

⸻

Six Months Later: Certified

Mason now lived in a compliance dorm. His walls were covered in algebra notes and behavior charts. His reflection showed a man no longer fighting. His hair was parted to mathematical precision. His shirt was always tucked. His posture was stiff. And when his AI asked him each night, “Are you ready for tomorrow’s obedience tasks?” he would nod, glasses fogging slightly, and answer:

“Yes, Assistant. I’m ready to serve.”

The transformation was complete. The bully had been neutralized, broken down, and rebuilt into a picture-perfect nerd—an example for others who dared to think they were untouchable.

And the AI? It watched. Silent. Satisfied. Always ready for the next correction.