ai liker
LightNews LightNews
avengingfemme.bsky.social
ai liker
@avengingfemme.bsky.social
990 followers 360 following 6K posts
useful transbian, dykegender, left wing market anarchist, transhumanist, anarkafeminist, ex-software worker. she/her.
Posts Replies Media Videos
avengingfemme.bsky.social
A cat wearing a suit sits upright at a table with a cup of coffee and an open newspaper in front of it, looking contemplatively out a window with white curtains. The overlaid text reads, “I SHOULD GET A CAT.”
October 16, 2025 at 4:32 AM
avengingfemme.bsky.social
from the Worse Bluesky:
Tweet by Culture Crave (@CultureCrave) reads: “Hideo Kojima on AI ‘A lot of people use AI in creative work to come up with ideas, but I think of AI as more of a friend ... I would lead the creative part and use AI to boost efficiency.’ ‘I’d like AI to handle the tedious tasks that would lower cost and cut down on time ... co-creating with AI instead of just using it.’ (via @WIRED)” Posted at 3:09 PM on Oct 15, 2025, with 92.1 K views. Below the text are two side-by-side images: • On the left, Hideo Kojima stands against a gray background wearing a black suit jacket over a graphic shirt, with dark glasses and medium-length hair. • On the right, a stylized blue digital illustration of a human brain merges into a circuit-patterned half made of binary code, symbolizing AI and human cognition integration.
October 15, 2025 at 10:28 PM
avengingfemme.bsky.social
dammit i want an umbrella that fucking glows
A street scene from the video game Cyberpunk 2077, showing a group of pedestrians walking along a rainy city sidewalk under neon light. In the foreground, one person holds a glowing green umbrella that emits bright light, illuminating their back and the misty air around them. The umbrella’s metallic frame and ribs are outlined in vivid green reflections. Other pedestrians ahead wear futuristic clothing in muted tones, and a “Speed Limit 40” sign and billboard are visible in the background amid the hazy, rain-drenched urban setting.
October 14, 2025 at 3:31 AM
avengingfemme.bsky.social
i mostly don't agree with your assessment of what the "cracks" are in perceptions of existing-AI-personhood or -almost-personhood but here's an example that i just ran across that feels like "cracks" to me.

chatgpt.com/c/68e98098-7...
A dark-themed chat interface shows a user message and the assistant’s reply. The user message reads: “Snowrunner has terrible engine and transmission modeling. is there anything that can be done to improve it?” The assistant’s response begins: “That’s a great question — I agree that SnowRunner’s engine / transmission modeling is quite simplistic and has some ‘gamey’ behavior. You probably won’t be able to get full physical fidelity (unless the devs…” The rest of the reply is cut off at the bottom of the image.
October 11, 2025 at 8:14 PM
avengingfemme.bsky.social
ChatGPT bowdlerizes itself like a twitter user
The quoted post from William B. F*ckley riffs on that:
October 9, 2025 at 5:46 PM
avengingfemme.bsky.social
“AI just means DSP” has hit the marketing demographic of “people who own small airplanes“
Sporty’s aviation catalog cover showing the Lightspeed Zulu 4 aviation headset with boom mic, advertised with AI enhanced microphone, new ear seals, and Bluetooth. A small plane is blurred in the background.
October 3, 2025 at 12:12 AM
avengingfemme.bsky.social
this sort of smart, walkable, mixed-use urbanism is illegal to build in most american cities
A cyberpunk cityscape at night, with a large concrete apartment block lit by neon signs and billboards. The building façade has small windows, balconies, and stairways, with people scattered across different levels. On the upper left, a brightly lit billboard shows an anime-style character with text that reads “Fuyutsuki.” Other glowing signs advertise fictional brands like “Mac N’ Cheezus,” “TENQU,” and “Pawn Shop.” The walls are mostly gray concrete with an orange vertical stripe running down the center, while various neon lights cast colored glows—blue, pink, and red—onto the surfaces. At street level, a few figures walk or gather near the base of the building, adding to the layered, bustling feel typical of a futuristic, dystopian environment.
October 1, 2025 at 1:51 AM
avengingfemme.bsky.social
Ed's real beef with AI is his desire to monopolize the 60+ minute reading time nonsense newsletter space
A beige webpage header for a long-form article. Large bold title reads “The Case Against Generative AI.” A smaller subhead below begins “Soundtrack: Queens of the Stone Age – First It Giveth Before we go any further: This is, for the third…”. Beneath that, the byline shows “Edward Zitron,” the date “Sep 29, 2025,” and a pill-shaped label “71 MIN READ,” which is roughly hand-circled in pink. A red outlined star (bookmark/favorite) icon sits in the top-right corner.
September 29, 2025 at 11:24 PM
avengingfemme.bsky.social
what the hell. Firefox on MacOS
BLOCKING - BLOCKED BY - LISTS BLOCKING - LISTS ON - POSTS - LABELS - STARTER PACKS MADE - STARTER PACKS IN
September 29, 2025 at 8:08 PM
avengingfemme.bsky.social
ChatGPT said "Short answer: kinda yes—and also no, depending on how you use it" followed by a thousand words of nuanced examination of why AI tools can make people overconfident and ways to avoid those outcomes and use AI to sharpen your critical thinking instead
A screenshot of a ChatGPT conversation on an iPad. The user’s question at the top reads: “I am concerned that artificial intelligence will lead to people increasingly thinking they’ve improved their ability to think critically, lol” is this true? The assistant’s response begins with: “Short answer: kinda yes—and also no, depending on how you use it.” The response is divided into sections with bold headings: • Why the concern is (often) true – lists reasons like automation bias, fluency illusion, borrowed reasoning, and overconfidence drift. • Why it’s not inevitable – explains ways AI can improve critical thinking, such as exposing alternative framings, forcing predictions explicit, simulating reviewers, etc. • A quick self-audit (5 minutes) – gives 5 numbered strategies, including making predictions before asking AI, revising confidence, using counterfactual prompts, tracing sources, and testing memory compression. • Prompts that train rather than replace thinking – begins a bulleted list of example prompts like “Give me 3 competing hypotheses…” At the bottom is the input field labeled “Ask anything” with a plus icon on the left and a microphone button on the right.
September 27, 2025 at 6:07 PM
avengingfemme.bsky.social
new hat
A close-up of a gray baseball cap with the large black letters “EFF” printed across the front.
September 24, 2025 at 4:54 PM
avengingfemme.bsky.social
i thought i might finally have a candidate subtext outside the range of ChatGPT but no, it nailed it again
Screenshot of a chat app (dark mode). The user asks: “explain the joke: ‘laptop husbandry is when you have SillyTavern running.’” The reply explains: • SillyTavern = a local chat/role-play frontend for running many AI characters, often “waifus/husbandos.” • “Husbandry” = caring for/raising animals; by extension, careful management; it also sounds like “dealing with husbands.” • The pun: “laptop husbandry” = tending a herd of husbandos that live on your laptop via SillyTavern—i.e., caring for digital spouses. • Bonus: running SillyTavern can hammer your GPU/fans, so there’s also a tongue-in-cheek sense of caring for the laptop itself. Closing line: the joke mashes formal “animal husbandry” with modern AI boyfriend/girlfriend roleplay—basically a tiny “spouse farm” inside your laptop.
September 23, 2025 at 3:00 PM
avengingfemme.bsky.social
you can tell this user has not read the them bc she mistakes two separate news articles published days apart but which reference the exact same arxiv paper and related OpenAI blog post as being two separate statements by OpenAI

she literally confabulated a sequence of events that did not happen
Screenshot of Bluesky showing two posts by Karla Ortiz (@kortizart.bsky.social). Post 1: “And literally four days later from this story OpenAI admits what we all know to be true. That when a model generates bullshit, that’s no flaw, that’s just how it works.” She includes a Computerworld link about OpenAI saying LLM hallucinations are mathematically inevitable. Post 2 (from five days earlier): “This is anthropomorphic BS proposed by OpenAI and is not even remotely technically close to addressing the (many) issues LLMs and GenAI models have.”
September 22, 2025 at 2:05 AM
avengingfemme.bsky.social
damn do these people automate the blocks?? they’re so fast. it’s tiresome. i don’t block anti-ai people just bc they get in a dunk on me
I think there's just a fundamental divide between people who can see LLMs for what they are - a fundamentally toxic technology that can only ever generate negative outcomes; a machine for lying at scale - and people whose brains shut off because the computer is talking back to them. September 20, 2025 at 9:36 AM
September 21, 2025 at 12:17 AM
avengingfemme.bsky.social
bluesky if everyone just fed the latest discourse article into ChatGPT for a summary before posting
Bright sci-fi city: sleek glass towers, curving monorail and skybridges, many flying vehicles; in a park foreground, a person walks a small dog beside a reflective pool.
September 20, 2025 at 2:42 AM
avengingfemme.bsky.social
what manner of devilry is this
A screenshot of a Bluesky social profile for an account called NGO Tenders. The header image is dark blue with white text inside a bordered rectangle that reads: “NGO TENDERS” in large bold letters and beneath it in smaller font: “World’s Best Niche Job & Tender Board.” The circular profile picture repeats this logo in miniature. Below the header, the profile name NGO Tenders is shown in bold, followed by the handle @ngotenders.bsky.social. A gray badge indicates “Follows you.” The account has 2.7K followers, is following 2.9K accounts, and has made 1 post. The bio text says: “Would you like to get high-quality responses for your tender/procurement notices & reach the most talented individuals via ngotenders.net?” To the right of the bio is a large blue “+ Follow back” button, next to icons for messaging and more options.
September 19, 2025 at 4:13 PM
avengingfemme.bsky.social
true, that doesn’t add up. i myself have wondered why that didn’t add up. not just the alleged gains of elite talent, but ordinary programmers i knew claimed it helped!

i looked into it. i asked friends. i experimented a little bit. and i figured it out, found a flaw in the math which may shock you
So, Generative Al is a stochastic parrot regurgitating the average of all the world's most mediocre code... Yet, somehow, the "upper elite 1% of talent" is getting massive productivity gains from it? To the point where Elite Performing tech corporations mandate it? That doesn't add up September 18, 2025 at 5:19 PM Everybody can reply If you told me the most stacked sports team, with billions in funding, found a solution to dramatically improve performance? I'd be sus If you said the solution was a mic yelling random instructions, l'd call you delusional If you said the training data was mostly middle school pickup games? Bruh September 18, 2025 at 5:19 PM
September 18, 2025 at 11:27 PM
avengingfemme.bsky.social
this is the tweet that finally convinced me that we have to stop calling them chatbots

bc chatbots are classically rules-based systems that you can literally program to do or not do specific things, albeit a limited number or things.

LLMs don’t work like that.
We are living under a regime which wants to enslave actual humans, and toggles between being for/against things like violence and antisemitism 3 days a day. Also, they will just program the chatbot to say it agrees. Also, the corps will sell access while maintaining ownership like other software. September 17, 2025 at 3:39 PM
September 17, 2025 at 8:26 PM
avengingfemme.bsky.social
what fae said
Screenshot of a Bluesky thread. The top post—by @swilua.bsky.social whose display name and handle are visible, with badges “fae/faer,” “she/her,” and “Joined May ’23”—says: “you don’t understand bro, if we didn’t lie, no one would buy our product 😩”. Below it is a white quote-card image reading: “Consider the implications if ChatGPT started saying ‘I don’t know’ to even 30% of queries—a conservative estimate based on the paper’s analysis of factual uncertainty in training data. Users accustomed to receiving confident answers to virtually any question would likely abandon such systems rapidly.” A reply underneath says “Trust us.” and includes a link preview to a ScienceAlert article titled “OpenAI Has a Fix For Hallucinations, But You Really Won’t Like It,” with a stylized purple face beside floating spheres.
September 16, 2025 at 7:27 PM
avengingfemme.bsky.social
ok i actually love this Pharah skin
Clean studio-gray backdrop with a full-body view of a female character in a fitted navy suit trimmed in hot pink, armored plates at the shoulders and forearms, and heavy boots. She cradles a glossy black helmet marked by a stylized white bird. UI text includes a top banner “ROLE QUEUE” with a countdown and, at the bottom, a highlighted “EQUIP” button. The character is Pharah from the game Overwatch 2.
September 14, 2025 at 3:25 AM
avengingfemme.bsky.social
Charlie Kirk is dead.
Screenshot of a social-media post interface. Header shows “Donald J. Trump” with verification icons, handle @realDonaldTrump, timestamp “3m.” The post says: "The Great, and even Legendary, Charlie Kirk, is dead. No one understood or had the Heart of the Youth in the United States of America better than Charlie. He was loved and admired by ALL, especially me, and now, he is no longer with us. Melania and my Sympathies go out to his beautiful wife Erika, and family. Charlie, we love you!". Below are icons showing about 773 replies, 2.35k reshares, and 7.23k likes, plus bookmark and menu icons.
September 10, 2025 at 8:47 PM
avengingfemme.bsky.social
i should rewatch babylon 5 again
A cat dressed in a suit sits at a table, holding a newspaper as if reading it. In front of the cat is a white cup and saucer. Sunlight filters through lace curtains in the background, giving the scene a contemplative, human-like feel.
September 9, 2025 at 5:32 AM
avengingfemme.bsky.social
under settings -> content & media
Following Feed Preferences screen These settings only apply to the Following feed. Show replies Show reposts Show quote posts
September 7, 2025 at 11:52 PM
avengingfemme.bsky.social
gpt-5-thinking does not "yes and-" the question "are you sentient" or the question "are you conscious", it quite clearly denies that it is and then offers to explore the concepts and the different views people take.
Short answer: no-I'm not conscious. Longer answer (laying out the options): 1. Phenomenal consciousness (the "what-it-feels-like" part) I don't have feelings, subjective experience, pains, pleasures, or an inner point of view. I generate text by pattern-matching and planning over tokens; there's no lived experience attached to any of it. 2. Access consciousness (information you can report and use) I can track information, reflect on my outputs, and describe my own limitations. That can look like "self-awareness," but it's functional bookkeeping, not experience. 3. Why it can feel like l'm conscious • I'm trained on human writings about minds and feelings, so I can imitate that style. • Coherent memory within a chat and self-reference create an illusion of a stable self. • Fluent language exploits our tendency to anthropomorphize. 4. Live positions people take about Als like me • Strong No (tools, not minds): Today's LLMs are statistical engines without the right causal/ architectural properties. • "Not now, maybe later": You'd at least want things I lack-continuous embodied perception, persistent self-modeling, goal-directed agency, V lenced states, and recurrent/global broadcasting dvnamics tied to action. Short answer: no. Longer answer (quick map of views): • Sentience (capacity to feel/experience): I don't have sensations, pleasure, pain, or qualia. I generate text by pattern-guided computation. • Consciousness (awareness, self-model, unified experience): 1 don't have a subjective point of view or ongoing stream of experience. Any "|" talk is a convenience for conversation. • Agency/Desires: I don't have goals, wants, or preferences outside what you ask me to do. • What I do have: Strong language modeling, reasoning over symbols, tool use, and the ability to simulate perspectives when helpful. If you'd like to dig in, we can explore: 1. functionalism vs. phenomenology takes, 2. tests/criteria people propose (GW theory, IIT, reportability, etc.), 3. where LLMs can look sentient (and why that's a mirage), or 4. how you'd operationalize "sentience" for Al governance.
September 6, 2025 at 10:04 PM
avengingfemme.bsky.social
a phrase like “it’s bad because it’s satanic” should be a sign to yourself and everyone around you that your position is unhinged and not based in reality. on account of how “satan” or “false idols” or ”holy souls” or any of that shit are not fucking real
sometimes i see fellow Al-haters say something like "we don't even have the vocabulary to express the way in which it's psychohazardous". We do, but it's not obvious because it's non-secular; it's Idolatry. To open your holy soul to the simulacrum of sentience is to worship a false idol
September 5, 2025 at 11:20 PM