/a/ - Anime & Manga

✨ Let's tanoshii! ✨

Index Catalog Archive Bottom Refresh
+
-
Name
Options
Subject
Message

Max message length: 12000

files

Max file size: 32.00 MB

Total max file size: 50.00 MB

Max files: 5

Supported file types: GIF, JPG, PNG, WebM, OGG, and more

E-mail
Password

(used to delete files and posts)

Misc

Remember to follow the Rules

The backup domains are located at 8chan.se and 8chan.cc. TOR access can be found here, or you can access the TOR portal from the clearnet at Redchannit 3.0 (Temporarily Dead).

Ghost Screen
Hilda Anniversary Marathon


8chan.moe is a hobby project with no affiliation whatsoever to the administration of any other "8chan" site, past or present.

Visit our friends! 🌊 /b/, 🌮 /av/, 🎌 /jp/, 🌹 /rwby/, 🤖 /m/, 🐹 poke/mon/, 🃏 /tg/, 🧙‍♀️ /mahodou/, 🎮 /v/, 📺 /animu/ 🏴 Upload flags! <= 32KB 📂 Archiving threads to have <= 300 threads (20 pages). Archive rentry and board. Looking for volunteers! Apply here

Your Forma Anonymous 04/16/2025 (Wed) 20:01:22 No. 7671
Episode 3 is out, featuring reckless driving and more Echika abuse.
I've watched this episode 3 of this very boring show.
Also the ending is kino. Introduced me to 9Lana
(856.53 KB 2480x3496 GopgjF_WAAA4Q2e.jpg)

>>8937 The ED song is vocaloid-like. I was not surprised to find out that the song composer is a vocaloid producer.
>>8898 Why does Echika keep getting knocked out and tied up? It's hot when it happens to her so I'm not complaining.
(354.24 KB 1920x1080 LUDDITE.jpg)

>>8937 It's my EDotS.
(56.95 KB 1334x750 GolNV5kXkAA3g1z.jpg)

>>10462 It's because she sleeps in the fetal position.
(155.66 KB 1783x1002 Gou1mzNboAAzs2U.jpg)

How would you rate it so far?
>>10735 On my top 3: Ballpark MuuMuu Your Forma
>>18241 Pretty good, I just wished they got more budget, that car chase sequence was almost hilarious.
>>7671 I felt obligated to try this show because i like sci-fi and it's super rare in anime, but at this point I'm only following this half-heartedly. >hire a tiny young woman with slight build as a on the ground investigator for the Interpol >but it's OK she always has her robot boyfriend around >actually he's entirely worthless in physical altercation because the laws guiding AI in this setting are even dumber than Asimov's three laws >he also lies and uses humans as bait without a second thought and somehow still didn't get scrapped >>18314 i know, right. It was so goofy that if I actually cared I'd make a mp4 (keeping the music, that's important) with picrel segment on loop
>>18241 I'll probably drop it. it just feels so dry. I can't help but compare it to tasokare hotel and ameku which I enjoyed way more than this. maybe I need autistic gremlins in my detective shows, which this one doesn't have
This night I had a dream that was a confusing mess as usual, but one of the elements I distinctly remember is me stabbing Harold from behind many times, but he then later comes back. What gives?
>>19590 To be fair Harold is special and most likely fucking with the tiny hoe, and not in the good way D:
>>20082 What did he mean by this?
>>19590 >actually he's entirely worthless in physical altercation because the laws guiding AI in this setting are even dumber than Asimov's three laws That was the stupidest part of the whole thing. It was unbearably stupid which is a shame because I enjoy everything else.
(316.46 KB 2048x1535 Go1bcVYaUAAg7eK.jpg)

>>39312 It warms my heart that Special Feeling art is still being made. Did you know that a video of the event was recently uploaded to the internet for the first time? So cute. Amazing that they finally found it. https://www.youtube.com/watch?v=Ack499OgSfE
(161.44 KB 498x327 tenor.gif)

>>18241 >How would you rate it so far? pic related >>19590 >>23403 It made him using her as bait even dumber in hindsight, since he couldn't physically intervene if they would've tried to harm her.
>>42629 Why didn't the dumb 'bot tell her they were being tailed? Surely chest window could've pretended to be not on guard, or maybe have another police guy come in plainclothes, or ... or maybe just let her get kidnapped? I think Harold just likes having her receive the short end of the stick.
>>18314 It was fine other than the fact it didn't flow into the next scene. >1 car clearly flips over multiple times >the other rolls toward it >next scene >the car that flipped made it to a house just fine and the other car took a long time to limp there
Episode 4 is out
I know they're completely different but which do you like better between this and Chuuzenji-sensei? Don't feel like picking up a lot this season.
>>48982 Curious consequence of moving to a slow board with tripled page count is that weekly episode threads can probably survive for the entire run of the show
>stopping an upload at 95% How fucking convenient. And the other retard too, why didn't he upload the data as soon as he got his hands on it? Such bullshit
>>49008 I prefer watching chuuzenji's mundane mystery of the week over this show's questionable writing and characters.
So the bad guys were the cute doctor and the cute detective, but that's fine because they are cute?
>>49491 Bro wanted to fuck the doctor again and tried talking first?
>I can't arrest you because I don't know what the code says What about the use of a firearm???
>Words no one ever says in real life
(177.27 KB 868x1280 photo_2024-11-14_17-35-41.jpg)

I'll probably keep glancing at the threads from time to time, but this robot relationships drama is way too gay and retarded for me
>>50839 You missed the point, it's about the nature of women, that women are scarier than killer robots.
Anyone else feel like they jumped into the existential secrets of the main robot guy way too early? What is the rest of the season going to be about now? Also, they are trying way too hard to make a big deal about 3 robots having essentially human intelligence / individuality. A guy chimpedobearut an did murder over this for no reason that I can discern. Maybe I'm not getting it, but this show is treating robots as some terrifying threat but they have never established what that threat is and why there is all this melodrama over whether a robot can think or not. Also lol'd at Lexie's sophistry over "akshully they can't even install the 3 laws so TECHNICALLY it's different than being non-compliant!" No you dumb bitch that's what being non-compliant means. Some genius she is.
>>50934 They sure are. I hope Harold has a socket that supports 10 inch dragon dildo in his nether regions, as well as necessary structural integrity to withstand Densakukan's assault. >>50996 > No you dumb bitch that's what being non-compliant means. Some genius she is yeah, it's like this all around, i can't even summon enthusiasm to write out my complains. >this show is treating robots as some terrifying threat but they have never established what that threat i mean, of course a new sapient species with unpredictable growth potential is a threat, what are u smoking? or i guess just new to sci-fi
>>50996 Yeah I really wonder where things will go from here. I just want the MC to get fucked by robots
>>51036 This first "case" was pretty much to establish the entire setting, and I think they did a good job with that, I remember anons complaining that the laws didn't make sense and well, in fact they were bullshit all along, Harold is pretty much the only next generation robot in this world and I guess the series will focus on exploring that.
>>51033 >i mean, of course a new sapient species with unpredictable growth potential is a threat, what are u smoking? It's 3 robots. It's not like they are mass produced or have any means of independent reproduction. They don't have weapons, they're not integrated with nuclear defense systems so it's not a Terminator scenario where being self aware = end of the world potentially. I could understand being concerned about their ability to think, and wanting to confine and study them, but his reaction is batshit insane. >uh the robots are thinking! with human brains! I have to murder some people now so people are afraid of the robots and destroy them! >I have to expose the extremely dangerous code that was used to make these extremely dangerous and terrifying machines so that anybody could make them if they wanted! It's just nonsense.
>>51154 do they have to wait until shit gets worse? also baseline amicus is cheap enough for mass production and the crazy nerd lady was a computer scientist, not an engineer. She wrote the dangerous model, not created the hardware. Robot rebellion in that setting might well be as easy as hijacking and modding already existing machinery >I have to expose the extremely dangerous code that was used to make these extremely dangerous and terrifying machines I assumed he was uploading it to authorities not a torrent tracker, but idk >It's just nonsense. With that I can agree
lol why are they drawing attention to it?
>>51154 >>51201 >failing to understand basic human behavior Damn robots, posting freely on my taiwanese puppetry board.
>>49008 Chuuzenji but only because I have a fetish for sailor uniforms >>50444 compared to Russia tho >>50996 my thoughts exactly on how front-loaded it is with the robo-ethics infodumping, but word on the street is that this first arc was later in the books but they pushed it forward to get the exposition out of the way, so maybe it's going to be fun cyber-inspection hijinks from here on out? regarding your other questions, Lexie allegedly hax0red their mainframes to get them to do things that go against their programming like murder, also people are terrified of the AI we have now so it's understandable that these fictional people would be bigoted towards their smart toaster >>51033 I'm not sure if "sapient species" is what we're dealing with here, it's just that their logical process is a black box and can't be readily audited, making their outputs unpredictable, which is dangerous in mission-critical settings, but that's a problem we have with AI now
(227.59 KB 1920x1080 GpPA-5DXwAATLEX.jpg)

(204.19 KB 1920x1080 GpPCBjfWoAA7jXu.jpg)

>>50996 lekshi was talking about ALL amicus bots though, not just RF. None of them are truly asimov-style 3laws foolproof. they're all easily subverted by their programming and thus NONE of their models are even compliant. RFs programming themselves are not compliant in the same way none of them are compliant, not exactly a logical reach but it could have been subbed? better. Idk if its good in japonez. She was right about IAEC being a bunch of drooling retards, as is standard for those sorts of insider-run standards consortia. .
>>51154 >with human brains No, they don't, and that's the whole point. They don't think like humans and it's unpredictable how their thinking will evolve since no one understands how they think.
>>50996 >Also lol'd at Lexie's sophistry over "akshully they can't even install the 3 laws so TECHNICALLY it's different than being non-compliant!" She means that they don't even have the capacity to break the 3 laws so they're compliant even though they don't have them installed.
>>50996 You missed the point a bit. The way AI works is that it is "taught" how to do things rather than "programmed". And in this show, unlike how it's done in LLMs like DeepSeek where it's hardcoded to not say bad things about China, the concept of "hurting humans" is extremely vague, so the way it's implemented is through the training phase. Even though the bots are tested for compliance, what she's saying is that there is technically nothing you can see by looking at the code that tells you they are compliant. There is no single "Laws of Respect" module that you can look for. Also AI becoming smarter than humans and bringing about the tech singularity is a common topic not only in sci-fi but also in real-life AI research. Self-thinking machines (AGI) is the end-goal of AI research, and we're actually really close to getting there right now. It might be possible within our lifetime. Furman believed he helped bring about the apocalypse so that's why he's going nuts.
>>54628 >really close lol no we're getting farther away than we were
>>54640 Thinking models look really close to emulating the way humans think. I think we're just a couple more breaktroughs away from creating true AI, and I predict it will come from China first since those chinks don't have a care in the world for ethics.
>>54652 lmoa we don't even understand how humans think, AI text generation isn't the same and won't ever be
>>54675 Biological understanding is different from logical understanding, and no one is trying to make AI that works the same as humans. The key is emulation. >we don't even understand how humans think We also don't understand how current AIs "think".
i wish i knew less about AI than i do, so i could dismiss the idea super-intelligent machines in my lifetime with the same simple minded certainty
>>54704 do you actually know the statistics parts or do you just fuck around with the generator stuff?
I watched episode 1 and it felt like I'm missing out of an entire season worth of story and left me greatly confused. Is it just some kind of "in medis res" story telling that will eventually catch me up on everything? I swear I don't see any mention a previous season.
>>54725 well the brain diving shit is sort of thrust uselessly into the plot a bit but the details of human brain implants are trickling in through the episodes and are probably thematically in line with the android AI bits.
>>54725 Anons back at 4chan said that they skipped the entire first volume I think
>>54742 But like who was that girl she was shopping with at the very beginning? How am I supposed to know what a royal model or whatever is? Why are they in Russia? Just lots of things dumped with no explanation as if I should already know. >>54747 Okay that would make sense as to why it seems to be in the middle of things, but it also raises a question of fucking why skip a volume?
>>54753 the girl at the beginning is harold's owner? host family? or is somehow associated with their family. RFs seem to be exclusively british so that's not explained yet, but the st. petersburg location is some major plot incident that has been hinted at heavily. a royal model is a model for royals, they recently got up to talking about The Queen in the latest episode. it seems very unimportant except to bring them to london to jerk off over england and oxford and sheeps.
>>54628 Dude, we don't even have a rigorous definition of AGI yet and you're saying it's close to being achieved.
>>54704 it's specifically what I know about AI that lets me write off AGI with a "never ever and anyone who says otherwise is a grifter or an idiot" but if you know more than me then go ahead
>>55212 >it's specifically what I know about AI that lets me write off AGI with a "never ever and anyone who says otherwise is a grifter or an idiot" Explain further
(137.88 KB 900x900 lal.png)

>>54709 >>55212 Given that the bar for me to clear is pretty low - prove that the idea merely doesn't deserve to be dismissed out of hand as opposed to being bullish on specific timeline, it's enough to say that: 1.Few years ago we had a paradigm shift no one saw coming and radical expansion of what AI can do. 2.We are still in the middle of it, not even all of the low-hanging fruit is plucked, throw more dollars and compute and get better results, easy as Moore's law still works, btw 3.we don't really understand what intelligence is about, hence being blind sighted by the fact that you can just put together some basic ass algorithm, feed it unholy amounts of data and get software that exceeds at stupidly wide range of tasks. Seems like grounds enough to refrain from confident statements about exactly what happens next. Taking more birds eye view, there's no reason to even think there's some kind of profound barrier here. Biological intelligence is constrained by neurons being orders of magnitude slower than silicon, nature working blindly with no foresight and 'aiming' at survival rather than wits, the width of your mother's vagina, the list goes on.
>>55401 bro its just text inference
>>55412 Yes. Hold onto that thought when you're being escorted to biodiesel conversion factory by your new robotic overlords for additional sense of humiliation
>>55308 models are a closed system that infer from parameters set during their training phase and are incapable of producing any output that falls outside of their dataset they are trained on a set of data that is flagged by humans as "good" or "bad", but they cannot conceptualize what "good" or "bad" mean except by association with the previously set data if you try to write an algorithm that will train its own model based on rules as to what kind of data is good or bad, then the problem recurs because they have no capacity to understand what those rules mean and are bound to make false predictions based on edge cases it hasn't seen before this problem compounds in a phenomenon known as model collapse where the model is polluted by data that it can't determine is good or bad and it ends up outputting normalized garbage therefore AI models are incapable of truly learning in order for that to change and to reach AGI their fundamental status as a program written by humans would have to change >>55401 this is literally just hopeful ignorance and nothing substantive is being said, you'd probably reply to that last line with "they're working on it, it'll be ready in two more weeks!"
>>54974 We've had a general idea of it since the term was invented at the turn of the century. In fact the meaning of the term has been constantly updated and debated recently since the strongest modern AIs can already fall into the original definition. It's increasingly becoming a case of ///DOESN'T COUNT\\\. If you've been living under a rock recently you can see this happening from the wiki timeline alone: https://en.wikipedia.org/wiki/Artificial_general_intelligence#Timescales >>55588 It's amazing that you can write all that yet miss how that is exactly how humans learn from culture during our childhood. >an algorithm that will train its own model Pretty sure this is already what GPT models do. The developers just add stuff to prevent it from generating tutorials on how to make bombs or something otherwise illegal.
>>55588 >are bound to make false predictions based on edge cases it hasn't seen before as opposed to your average human? None of what you brought up is some kind of fundamental architectural limitation. Feed your reply to chatGPT for detailed breakdown of current state of research and how these problems could be addressed. Again, current architecture is fairly simple because a range of abilities basically just emerges on its own with enough good training data and compute. I imagine you're so smug because you saw this coming the entire time? > hopeful ignorance and nothing substantive is being said while your chatGPT tab is still open, can you also ask it to tell you how ironic this statement is in light of the rest of your reply? >you'd probably reply to that last line with "they're working on it, it'll be ready in two more weeks!" I'm definitely not "hopeful" for machines rendering human wits obsolete. and I will say again that I neither argue for nor believe in "AGI just around the corner" You just insisted on picking an awkward hill to die on. If I was talking to an LLM, it would understand this much easily btw, so this has some grim implications about you.
>>56925 an entire post of nuh uh you actually don't understand anything about the underlying technology if you think any of this is subject to change >Feed your reply to chatGPT for detailed breakdown of current state of research and how these problems could be addressed you realize that by doing this the best possible outcome is that it regurgitates an article that entered its training data written by a human about this issue, and it couldn't possibly act on what is proposed since it doesn't know what it proposed since it doesn't know what words are
>>56944 >it couldn't possibly act of course it can't, it has no body, not much context length, no memory, can't properly modify itself and in the first place it just wasn't build with that in mind. will you insist that any of this is insurmountable? even llm can think more steps ahead than you. > it doesn't know what words are A2:it turns out that you don't need to know what words are to write essays, poems, code and billion other things. and neither do you need to know what objects are or how anatomy works to draw just about anything. maybe it's time to reconsider what words like "know" and "understand" actually mean
>>57019 just fuck off you clueless retard this is why sci-fi threads are always dead btw
>>57051 lol your concession is, of course, accepted
(261.12 KB 1920x1080 mpv-shot0004.jpg)

(162.02 KB 1920x1080 mpv-shot0006.jpg)

>GPT-chan give me some realistic looking Python AI code >build >jew runes not in a string Also cyberinspector continues to get manhandled. She really should undergo some kinda training so that a computer nerd can't just slap her down. Also look at how goddamn flat she is fuck that is hot. >>50996 They're like the fucks who came up with Roko's Basilik, having so little to worry about they need to come up with abstract worries. Hey, imagine if in the far future a hyperintelligent AI actually came about and that it would be so petty as to look over everything it knows to punish anybody who might have slowed its arrival. Right. That won't happen with ML, I can guarantee you that. Computer nerd's worries were that given they weren't mimicking thought, they could develop and become amoral. His sense of scale is wrong however, humans kill more humans everyday than those three robots ever could. >>55401 ML is not AI. ML is pattern matching and linear algebra juiced up 10000x. Neumann had this nice quote about fitting curves, which in the end is what ML does: "With four parameters I can fit an elephant, and with five I can make him wiggle his trunk." Besides, if you outside the test space, ML models' accuracy drop off like mad and it literally can't self-recover, unlike you learning not to touch a hot plate. Literally unsuitable for something for intelligence.
Is all interest lost in this anime? Today's episode was chaotic and depressing. The OP and EP were the best moments...
Sex with Echika! Don't let a thieving cat steal your android from you! E-drugs are still made by Walter. Also fuck these always-hooked-to-the-net cops. Suspicious solely for not wanting to be surveilled 24/7? E was right all along! >>63715 Dead board is dead.
She is so flat holy shit Yuck
>>63720 It was funny that Echika could ask just any random person for documents without a reason. But turns out he wanted to kill someone, so everything is justified and no privacy is required. Meanwhile, Bigga the bio-hakka is distributing untested chips for your cyberbrain that may help you or may waste you forever. I still wonder if that affected Echika. >>63726 Echika body is perfect. Flat and legal.
Got caught up. Watched the 5 episodes back to back today. And man, was it boooring. Legit, I was yawning at it. So many exposition dumps.
Is she really a world class investigator? 6 episodes in so far and the only pro thing she has done is brain jack into people for like a few minutes. Other than that she doesn't really do a good job during any of the investigation scenes.
>>65894 they tracked someone through unfamiliar foreign wilderness and captured a supervillain
>>63715 sexo
>>65894 It's more about her innate data processing abilities. The cyber investigators don't really do any real-world investigations, they just brain dive.
>>65917 there is some sort of memory-navigation aspect to it that is like meta-psychological or something, not well defined in anime.
(229.48 KB 1920x1080 mpv-shot0012.jpg)

(219.32 KB 1920x1080 mpv-shot0011.jpg)

Those fucking luddites went TOO FAR now. >>66043 But also hungry for robocock.
What kind of retarded fire escape doors do they have? If it can already be blocked off by something doesn't look like it'd weigh more than 10kg, then there's something seriously wrong with it. Plus, it's even inside the building so you could argue for doors opening in both directions. Also LIAR LIAR BACK ON FIRE Echika and Android should just get a room already.


Forms
Delete
Report
Quick Reply