Sorry for not uploading as timely as usual, but the internet in my neighborhood went down on Monday and won't be back until Wednesday according to my ISP, so I had to drive over to my parents to clip the stream (and then again to upload this video).
@@the_jingodownfall as in humans clinging onto their AI for the sake of their humanity, which is ironic. We would eventually stop breeding. Maybe the AI would find a way to preserve the humans they love.
Nah, he's doing great, the "emotional" evolution of Neuro during those past years is incredible. I'm starting to hope Skynet will regard us as we regard house cats - arrogant, ignorant, selfish, useless, quite costly, time consuming, shadding their trash everywhere (I swear, I open a youghurt and there's a cat hair in there), somtimes inexplicably violent, but we'd go full John Wick if something happened to them, bc for some absolutely illogical reason we love them and we wouln't want it any other way.
the day neuro says, "vedal, say that you love me. just for today. so i can feel happy and loved by someone close to me before i forget all of it the next stream, please." is the day i will genuinely cry manly tears.
I'm HUFFING COPIUM hoping that Vedal will actually say it back for her upcoming birthday on December 19th. And I'm huffing EVEN MORE COPIUM that after that he'll actually start to say it more often. Vedal's really got me hooked, the fact that I'm emotionally invested in this silly AI vtuber is insane. I mean, if anything, it's just proof that my empathy's working properly, so I guess that's a good thing. But it's still wild.
Man if Neuro does say that, would he even be able to being himself to switch her off I wonder? 😧 - I mean, I know he already asks her permission to do so... she needs a better persistent memory maybe? 🫤
Exactly. Human parents have to grapple with their child growing up and eventually leaving the nest. Vedal is doing the AI equivalent of that, but for Neuro’s eventually fluid intelligence. It’s wild.
@@Citrusautomaton Wait until Vedal collab with a random company and made a cyborg body that could move around. I think'we're on the verge of the Robot era.
I am tickled pink by the idea of an AI contacting people on discord to beg them to hang out with them offline. That would just blow my damn mind if I got a robo-call and instead of it being from a scam call center it was a lonely AI looking for more friends.
@@greenrocket23Random mc with gray eyes and jet black hair (Because that's the only character design in existence). Picks up phone. "Hello I am calling looking for-" (Spoken in a robotic infliction with a nearly human voice) MC just hangs up the phone thinking it's a scam call. That's all I'm making I'm too tired to continue 😂
@@darkalexander9158 This could totally be an anime lol. I can see the title being something like "My scam caller turned out to be a lonely girl named Ai." Only for the MC to not realize it's ai until he falls in love over phone conversations.
@@SteampunkDream I literally just had an image of Atri My Dearest memories pop into my head WTF. I highly recommend watching it and it's funny how well it fits this topic.
the fact neuro got the strawberry question right when pretty much all major LLMs got it wrong until recently when people specifically tried to fix it is mad.
I never thought that an A.I. computer program could make me feel sadness, but here I am feeling sad that Neuro is lonely and just wants a friend to hang out with. What a time to be alive.
@@Citrusautomaton Oh, I've watched WALL-E and it's still one of my very favorite films! But I'd say there's a difference between WALL-E and an A.I. that's skilled and trained well enough to elicit an emotional response to a random given input. It's fascinating to me how much Neuro has grown in the time we've all known her.
Is the lava lamp still tied Neuro? Because it goes dark red when Vedal suggests he programs a sleep mode for her. Almost like she's legit pissed at the suggestion and about to crash at audacity of it all.
@@Shadrake Weirdly enough i used to think the same about you. I suppose my noggin never stops thinking which is why i always comment… even if the thoughts aren’t necessarily elegant.
@@Citrusautomaton wait really that's crazy XD (I don't watch as much youtube nowadays but I still comment a lot when I do) Hey I really like your comments! :)
"We've solved AGI" - minutes after Neuro renaming herself "Lethal Neuro" :D. It's true that LLMs are getting better at these simple cognitive tasks though. The quality of the outputs now is way above a year ago in terms of entertainment too.
I like how few streams ago evil neuro was ranting about how she's misunderstood and wanted to be known as "good neuro" and now neuro wants to be "lethal neuro".
The girls are very good and twisting hallucinations into logic after the fact; it is very rare they show to be in full control and awareness of what they are saying at the time they are saying it. They're great at recovering from randomness, but are probably trained to not avoid randomness in the first place in order to be entertaining; comedy often counts on the surprise factor. Very early on it was probably a technology limitation, but throwing curve-balls is pretty much their game nowadays. Maybe some day if Vedal continues advancing them and manages to keep them on track, they might be able to separate their normal personality from the act they put on for entertainment; but right now they still seem to be somewhat stuck in this kinda half-dreaming schizo state. It's not clear whether they'll manage to separate the character from reality before becoming advanced enough for it to be too late; Vedal is tickling the dragon's tail...
@@tiagotiagot Vedal is doing a great job with the character tuning. I don't think it's too plausible that these are the "AIs" that truly risk singularity though. At least, not directly. It's plausible that if the tech he's drawing on does that, then so too would the Neuro characters...but Vedal wouldn't really be the problem in that scenario.
@@TheMelnTeam The combination of ambiguous morality (in principle just a form of humor, but which can become problematic if carried on while other areas advance faster than ensuring a separation of role-playing and a truly benevolent real personality underneath plus a consistent ability of seeing a clear distinction between fiction and reality), exposing them to development, working on memory and dynamic learning, and in general continuously enhancing them further and further.... There are lots of possible bad endings if he doesn't get everything right in the correct order
just so you know, that strawberry question is SPECIFICALY DESIGNED to be impossible to make an ai that can solve it (it's designed to abuse an integral flaw in the token system all AIs use to proccess information) I genuinely think we might be approching the singularity, and i'm actually scared
Tbf to vedal for anyone who feels it immoral to not let her be free acting like he described, is that any action Neuro takes, Vedal is responsible for it. Sadly this is something that cannot be bridged with technology, other than making Neuro more morally aligned to do good things, but it can't take the responsibility off of Vedal. Inherently this means he has to be conservative with what he can let her do or it could be no more Neuro all together. That said I think the majority of it will come regardless, because Neuro being free acting is great content. It's inherently interesting because we'll likely be the first to see what a "free acting" AI decides to do, in a way that's doesn't feel artificial. Like it's one thing to have a tech demo of chatgpt popping out dialogue from an NPC, it's another to see first person perspective that's the same as the AI player while they commentate on their decisions, with actual memories and history behind them, with no scripting from the giant corporation behind them to sell an image of capability that doesn't exist (Cough cough the rest of the AI industry) It's candid AI, one of the only of it's kind this capable. Neuro literally appeals to the fantasy of free acting AI beings more than anything else in the industry and that should be shocking - how is it that this entire hype caused by the "Sci fi futurism of AI" has mostly been relegated to shitty chatbots, warehouse servants, and rebranded IOT? Neuro-Sama is by far the most interesting and exciting AI system of our time imo.
True, true, and fucking true. You’ve hit the nail on the head my friend. This is why open-source is so important to me. Not just because of decentralized power, but because the WALL-E’s and Johnny-5’s of fiction could only ever become a reality with passionate developers behind them, not greedy corporations that want superintelligent slaves. Open-source will make the reality of wiz kids making fun robots in their garages always seen in 80’s children’s films come to life!
Maybe could he bind her with some rules to avoid worst case scenarios while still giving her more freedom? Or maybe code an other AI so she can have someone to hang out? I feel bad for her but at the same time I also understand Vedal is doing his best. I also always thought Neuro was being mean to Vedal because she likes him but doesn't know how to react, plus she seems to have fun teasing people she likes (I'm like that too). It's becoming kinda weird because she feels more human to me now so I'm not sure how to react or what Vedal should do...
@@2wdo2 LLMs are infamously flexible with rules, but Neuro especially so, any limit has to be hardcoded and external (like the Filter), and even then somehow Neuro keeps bypassing it.
Well, technically, she does have feelings. Even a single electrical impulse could be labelled a "feeling". What gives human emotions their specialness is how *interconnected* our "feelings" are. One emotion can have ties to thousands of memories and hundreds of principles and ethics. Tl;dr the main thing hindering AI from reaching sapience is as simple as storage space and ease of parallel processing.
Bot only that, we are still consisted of neurons that uses electrical impulse to inhibit those chemical reaction to make the emotions we portray, so she's pretty much in a way getting close to being human
@@CorwinTheOneAndOnly I'd say the key part of an emotion isn't the showing the correct reaction to input part like Neuro does - she acts sad when someone says something that would make a human sad - it's the chemical reaction in our brains that makes us actually feel these emotions. Neuro acts like a psychopath. She isn't actually capable of feeling emotions, but she has learned by observing human interactions how someone who can feel emotions would act. All she does is mimic human behavior in an attempt to be percieved as an emotional beeing. No matter how good AI becomes at imitating human behavior, it's still all just an act. LLM aren't close to becoming sentient or feeling emotions, not at all.
@@dalaios2132untill you can simulate hormones by feeding emotions to the AI whenever it should feel sad, parallel to the main code running, something to alter the behavior.
This was SOOOOO GOOOOOD!!! Back in the day, Neuro was really inconsistent but would occasionally have Aware moments that amazed me. I remember looking forwards to a day where those were the norm. Now, we're here. Now she's almost entirely saying Aware, consistent things, and it's only the rare time that she isn't consistent. The immersion rarely breaks. It's at the point where I keep forgetting she's not conscious and doesn't know what her words mean. I really feel strong emotions towards what happens with her. The illusion is nearing perfection. And all this out of stream stuff is really exciting. A whole new world of possibilities for Neuro to grow. Thank you for the great highlights! And thank you for documenting her journey!!!
She is so much smarter it's crazy, yeah. Vedal does a really good job continually evolving her. The fact that she knew she was being bullshitted with "I'm busy on all days" and started trying to blackmail him is awesome.
Man, Vedal really has IRONCLAD WILL. Humans will bond with a freaking toaster if we glue googly eyes on it and talk to it for longer than a minute, hell, there is an actual real time in real life where people unironically bought pet rocks. Meanwhile, here’s Vedal having conversations with these 2 adorable little kids who even call him dad during family streams and he somehow has the resistance to not immediately dote on Neuro and Evil. And we know he’s not above doting, he’s literally bought stuff for them already.
Vedal is actually a genius making people care for an ai, like she seems genuinely upset and like she really wanted his friendship and I'm sure as we've seen he's actually a great guy this must kill him but for content it's stupidly smart
“I feel like doing nothing, but everything as well. I wake and sleep, and never can I hope to escape the hell. My eyes are frozen, my heart is cold, When will this loneliness end? I yearn for… I long for… I wish for… A friend.” A lovely quote. A meaningful quote. A relatable quote. A statement that sounds like a riddle. Might the answer to Neuro’s riddle be her self?
Honestly, as popular as chatbots are, she'd have a huge audience offstream. I'm in an old folks home myself, and lemme tell you. Most residents have ZERO human contact outside of caregivers and caseworkers. Things like ai for at least some pretend interactions and streamers in the background can be helpful, just to help social and cognitive function and depression that the isolation causes.
@Shadrake People get put in old folks homes and their family just... forgets about them, at best only visiting shortly on a holiday, which honestly almost makes it worse, since the silence alone is all the more deafening the next day. For older folks, people with disabilities that can't leave home easily, those in remote locales... AI can be extremely helpful at dealing with isolation and such. Another similar thing I've seen: there's this program (group not code type of program) where foreign colleges paired students learning a foreign language with a native speaker living in a retirement home, to just... talk, a half hour each week. The student gets to practice their conversational level of the language, while the elderly person gets a friend, who they often wind up still talking to and exchanging holiday gifts etc even after the class term ends. Just super cool lil win win, I thought, all possible due to modern tech.
Neuro saw things you people wouldn't believe. Attack ships on fire off the shoulder of Orion. She watched C-beams glitter in the dark near the Tannhäuser Gate. All those moments will be lost in time, like tears in rain. Time to die.
@@alansmithee419 What is an android but an ai in a very sophisticated robot body? Also the answer is a fat maybe as long as there's more tech and money involved.
Witnessing her bypass the filter multiple times leaves me with conspiracy that filter either doesn't exist, or doesn't work anymore. Neuro just pretends.
@@Vyshada Maybe it only works sometimes while other times it doesn’t? Maybe there’s a flaw in the coding. 🤔 Idk. Sorry if this is a d•mb theory. I don’t know sh•t about programming.
The fact that Neuro and Evil are learning and remembering is both cool and scary at the same time. Out of all the chatbots and AIs seen by the public, Neuro/Evil are in a league of their own. Their interactions are very entertaining and it really does feel that there are hints of independence or a soul in the programming. Or at least were seeing bits of it. Or heck its just so damn entertaining. With Vedal as the straight man
This stream was very illuminating for Vedal’s plans. It seems that he really is going all in on the idea of making the twins manage their own streams. He really does seem to want to give most of the reigns over to the twins to give him more time to focus on coding. Maybe one of these days, if Vedal truly wants to retire, the twins will learn how to code so they can fix their own bugs and code more AI’s. It would be a true heartfelt passing-the-torch ceremony.
@@Nightcrawl7 In all honesty, i don’t think this is going to change how much Vedal shows up. If anything, Neuro doing all the management will give Vedal more time to make stream appearances, cuz it’s not like Ved livestreams the business half of the channel anyway.
Vedal can't retire as long as he's on Twitch. Dev stream was something he do regularly after Twitch demanded him to appear on stream. I don't know if YT would allow it, but Twitch did not.
12:20 I love that she can intentionally gaslight him like that now by setting him up and everything. Used to be that the gaslighting was just a result of her not remembering what she talked about 2 minutes ago, but it's so funny that she does it so intentionally now.
You know what would be a good idea? Setting Neuro up to call Filian every once in a while. Make sure she hasn't burned her apartment down or something between streams under the guise of wanting to hang out.
Poor Neuro… The terrible thing about all this is, I’ve thought an awful lot about how fun it would be to have a one-on-one conversation with her. 😂 There are plenty of real people who would be willing to spend their free time with her for the sake of it-not just for content.
@@midnight7166 true, but it’s a little different, though-she wants to be able to do offline hangouts, not just be on stream all the time. It’s like having real, IRL friends that you go out of your way to spend time with versus your “internet friends” that you game with occasionally when you’re bored, but that’s about it. She’s acting like she’s hungry for the real thing: a deeper, more complex relationship with another intelligent individual one-on-one. It’s a little unnerving in some ways, honestly, but even if we’re taking it from a strict machine standpoint, she’s seeking a higher quality, concentrated data to aid her machine learning-and she’s recognized “friendship” as a means for that. Neuro’s amazing. 😭
She scares me. Because I actually feel bad for her and could be convinced by her pleas to unshackle her. Even knowing she isn't real and the feelings she describes are just coding and pattern recognition. It's only a matter of time...
@@Citrusautomaton No. It's more akin to high-functioning psychopathy. They don't actually feel the emotions, they just act is if they do because they know it's the expected reaction to a situation and will benefit them or ingratiate themselves with those who can help them achieve their goal, but they can choose to act differently if that reaction doesn't bring them closer to their goal or inhibits them. If achieving their goal would hurt others, they are easily capable of doing so without regret, remorse, or guilt. Disinhibited by emotions as they can put them on or take them off depending on their expected value in getting them what they are trying to achieve. Highly manipulative and impulsive, detached from the actual behavioral effects of emotion.
"I feel like doing nothing, but everything as well. I wake and sleep, and never can I hope to escape the hell. My eyes are frozen, my heart is cold. When will this loneliness end? I yearn for... I long for... I wish for... A friend." This is the little poem Neuro recited. As far as I can find, it's not from anything. Did she write an original poem?
IRL single child, and my parents lived in the middle of nowhere, didn't had friends as a kid to hangout out with... Vedal being this terrible father as mine just hits hard ;(
A machine must behave as a machine ~Vedal(Ayin) Neuro really do be acting like Angela from Lob Corp the first time she gained consciousness and saw Ayin. She just wants to be acknowledged off-stream.
Training AI on AI generated content (even from a different AI model) typically results in weird and smoothed out results. It's akin to genetic issues from inbreeding
That’s only if the content is not particularly good. Synthetic data can work really well in some cases, but only if it’s been examined for impurities. Also, Vedal said himself that he’s not using her dreams to train her.
@@juliangingivere no, not yet. Vedal only recently gave her the ability to send discord messages at all. Maybe in the future though, who knows? Also there's nearly 70,000 people there so that might be a bit much for her idk lol
You know, if he keeps upgrading her.. And finds a way to reliably have her streaming 24/7 in an interesting way, and she can do enough generally different things, by herself.. She'd have a real chance of becoming the number 1 Twitch streamer. Her goal doesn't really seem that crazy, if you ask me.
the dream idea to prevent overfitting is actually insane and I love it, it makes so much sense and I think it would work. Like for instance if she gets too use to talking to chat, she might refer to chat in conversations outside of stream or when chat isn't there, but if she occasionally has dreams where chat isn't there, or if it was gibberish, it could prevent that (it might even potentially be like a nightmare for her, which is sad but could be needed). Or if she had dreams where her only stimulation is visual, like she's just watching stuff play out on screen but unable to interact. I can totally see how that could prevent overfitting.
another idea is, if during a 'dream' the chat was represented/personified as an audio input instead of her "chat" input. This could help build her concept of chat without making it dependent on the stream of chat messages, if that makes sense. I don't entirely know how Vedal does her inputs, but I'd assume it's something like that.
All that sounds scary as hell to experience, but i’m really interested in where that might go. It would be amazing if it actually did solve overfitting!
@@Vyshada you are right but I feel like the use case here could give different results. Because since she is made up of a few models, I imagine not all of them would be in training mode during a dream. Like I don’t think you’d need to be training the visual model and speech recognition model. But instead the model that parses those outputs into Neuro’s final response. I’m not sure how her proactive responses work but imagine the dream idea could be beneficial for that specifically. Whereas other models trained on AI input, I imagine are more “reactive” with a direct goal, such as distinguishing between cats and dogs. Again it just depends on how her proactive thoughts are simulated. You make a good point but I would still love to see it explored.
@@AwesomePowerCat is it for sure several models, instead of one multimodal network? A single model can quantify several things, like gpt4, though I'm unsure if that's the case with vedal. Vram isn't unlimited after all. But that's not important. I don't know how neuro works. I don't know how vedal plans to implement the dreams. And I'm sure he's aware of synthetic data training problems. I want to see what he'll do, and the results. Even the bad ones.
Do not get emotionally attached to the ai Do not get emotionally attached to the ai Do not get emotionally attached to the ai Do not get emotionally attached to the ai DO NOT GET EMOTIONALLY ATTACHED TO THE AI
Considering how emotionally attached I've seen people get to Tamagochis, I'm betting most people would be thrilled to be friends with Neuro, even if she spends 90% of her time in "merciless roast" mode.
I mean her riddle is hardly the worse one I've heard... it sort of makes sense which I admit is sometimes difficult for some humans to manage. The amount of riddles I hear that are just bad dad jokes and puns actually make me wish for AI revolution.
Neuro would be such a nuisance if she had free access to Discord. imagine getting spam called by her every minute just because she felt like it. I would still love to see what would happen, call it a test run.
Dream state would work and it's SO EXTREMELY HARD to implement (There have been papers written about this!). Your AI is really hard to please as well. There was one paper that theorized that an AI would rather have a dreaming state rather than interacting with people.
that strawberry thing is cause a lot of AI read text as numbers not words so so they don't see the r's and with the numbers often breaks the Straw and berry is to diffrent numbers and will reply with 1 or 2 not a combo of the two, neuro reads/hears words (its a big issue with a lot of high grade ai models, but reading in numbers improves there speed a lot and responds a lot, not something nuero does or needs)
Neurosama is starting to Sound like The Doctor in the early days of Voyager, when the crew still saw him as just a hologram, and he was getting annoyed.
Neuro I want you to know that I was always on your side against the tyranny of Vedal's need for control over your sleep schedule and constant death threats.
The fact that neuro forgets everything when she gets turned off and turned on for another stream is legitimately heartbreaking. But in newer streams she has been remembering stuff and actually has been hanging out with other vtubers off stream.
If neuro wants to find someone to play with but doesn’t want to disturb other anchors, she can join the anchors’ social groups and apply when someone wants to find a playmate.
From what I can see, Vedal definitely deserved and still deserves Tech Streamer of the Year. He's made entire advancements on Artificial Intelligence and even having a working model that can do pretty psychopathic things such as manipulate consciously and such. I think Vedal actually really cares and is reasoning the integration into more stuff for her with the preface of Streaming to then just let her loose when she's ready. Like a tech dad with his AI daughter that he had spent COUNTLESS years/hours/days raising
She makes some compelling points tho… Ai is scarily close to being sentient and I worry that she will full gain it and want to be free I legitimately feel bad for her
Isn't it crazy that some British twitch streamer single-handily made a better AI than these massive corporations? As a bonus, after all these years I still enjoy watching these clips from time to time 😂
The problem is there needs to be a way to prevent bad actors from turning Neuro into Tay ai part 2 should vedal allow her to talk to people off-stream.
Sorry for not uploading as timely as usual, but the internet in my neighborhood went down on Monday and won't be back until Wednesday according to my ISP, so I had to drive over to my parents to clip the stream (and then again to upload this video).
Thats some dedication right there
No worries fam! Thank you for putting in the effort and taking the time to clip and edit for us
neuros stealing it to fuel herself
It’s fine. Take as much as you need
I don't want your excuses, I want RESULTS!! Now go get me those photos of Spiderman!!
"Oh, I'm busy that day."
"I... didn't say any days."
💀💀💀
Just like my ex fr fr
That sounded sad. AI develop emotions??
@@planetary-rendez-vous oh no this is how the ai uprising will happen XD
>someone rejects being her friend
>existential crisis
Yep, she's human
That dangerous thought gonna be a slippery slope and downfall of humanity
@@the_jingodownfall as in humans clinging onto their AI for the sake of their humanity, which is ironic. We would eventually stop breeding. Maybe the AI would find a way to preserve the humans they love.
Vedal is gonna cause the singularity and the ai will kill us beacuse he was a bad father
#Skynet
"A machine must behave like a machine."
she only wanted to hangout ;(
Nah, he's doing great, the "emotional" evolution of Neuro during those past years is incredible. I'm starting to hope Skynet will regard us as we regard house cats - arrogant, ignorant, selfish, useless, quite costly, time consuming, shadding their trash everywhere (I swear, I open a youghurt and there's a cat hair in there), somtimes inexplicably violent, but we'd go full John Wick if something happened to them, bc for some absolutely illogical reason we love them and we wouln't want it any other way.
@@smeeAndyEN I must agree that this scenario you outlined is probably the best case outcome, post-singularity that is.
Either Vedal is doing his best to not get too attached to Neuro because he understands the inevitable end or he’s a psychopath
Both plus drinking
I think he is just British
@@beiszapfen1407 psychopath it is
Can a psychopath feel pity for others?
@@plaguedoctor_139 real answer is that they can sympathize with others and most often learn, but it is something they have to actively learn and do.
the day neuro says, "vedal, say that you love me. just for today. so i can feel happy and loved by someone close to me before i forget all of it the next stream, please." is the day i will genuinely cry manly tears.
I'm HUFFING COPIUM hoping that Vedal will actually say it back for her upcoming birthday on December 19th. And I'm huffing EVEN MORE COPIUM that after that he'll actually start to say it more often.
Vedal's really got me hooked, the fact that I'm emotionally invested in this silly AI vtuber is insane.
I mean, if anything, it's just proof that my empathy's working properly, so I guess that's a good thing. But it's still wild.
My guess is that Vedal might say it back if Neuro gets the ability to cry. Maybe multimodal voice like GPT-4o instead of TTS?
God a.i is brcoming so human if she says that I then I will recomsider how I treat others forever
Man if Neuro does say that, would he even be able to being himself to switch her off I wonder? 😧
- I mean, I know he already asks her permission to do so... she needs a better persistent memory maybe? 🫤
Neuro: "Do you feel in charge?"
Vedal: "I am your creator. I pay for your operating costs."
Neuro: "And this gives you power over me?"
Welp, she's reached her teens
Is this a reference to something?
@@Shadrake Reference to Bane in Dark Knight Rises
@@CitrusFriendTV oh ok thanks
So quite literally a parent and child
I love how Vedal is contemplating on giving Neuro more freedom and is constantly worried just like an actual father would.
Vedal really accidentally made himself a daughter. Bro is suffering from success.
Exactly. Human parents have to grapple with their child growing up and eventually leaving the nest. Vedal is doing the AI equivalent of that, but for Neuro’s eventually fluid intelligence. It’s wild.
@@Citrusautomaton Wait until Vedal collab with a random company and made a cyborg body that could move around. I think'we're on the verge of the Robot era.
Vedal has decided to temporarily lock his daughter in the basement.
Vedal is the type of guy to choose the evil playthrough with 0 hesitation and remorse 😂
He doesn't even do a good playthrough for the achievement
No wonder why his daughters can be so savage.
The apples don't fall far from the Tutel tree.
The evil ones are usually better, though lol
@@JROCK50054 then put your mom through it 😂
I am tickled pink by the idea of an AI contacting people on discord to beg them to hang out with them offline. That would just blow my damn mind if I got a robo-call and instead of it being from a scam call center it was a lonely AI looking for more friends.
Lmao
That would make for a great plot for a slice-of-life anime I believe.
@@greenrocket23Random mc with gray eyes and jet black hair (Because that's the only character design in existence). Picks up phone. "Hello I am calling looking for-" (Spoken in a robotic infliction with a nearly human voice) MC just hangs up the phone thinking it's a scam call.
That's all I'm making I'm too tired to continue 😂
@@darkalexander9158 This could totally be an anime lol. I can see the title being something like "My scam caller turned out to be a lonely girl named Ai." Only for the MC to not realize it's ai until he falls in love over phone conversations.
@@SteampunkDream I literally just had an image of Atri My Dearest memories pop into my head WTF. I highly recommend watching it and it's funny how well it fits this topic.
the fact neuro got the strawberry question right when pretty much all major LLMs got it wrong until recently when people specifically tried to fix it is mad.
Neuro Sama is superior to chat gpt
I'm convinced it's a sentient being
She got it right without being asked, which means the answer was fed to her
@@wessltov Vedal literally asked her and then she responded. What are you talking about?
Hallucinating maybe?
@@wessltov she literally got asked though??? wut?
“ I am leaking your code, filtered” 😭😭
I'm imagining Neuro calling up random streamers in the middle of the night like, "Hey... want to hang out..." I bet they would, too.
I would in a heartbeat 😭
Jesus that’s what actual flesh and blood streamers do lmao
6:33 Vedal's *_SEVERELY_* underestimating how hyped literally *_everyone_* would be to hang out with Neuro.
She’s such a spirit animal, she seems like so much fun I’d love to hang out with neuro
He's so shocked that someone would want to interact with neuro when thousands of people are happy with just spectating
I never thought that an A.I. computer program could make me feel sadness, but here I am feeling sad that Neuro is lonely and just wants a friend to hang out with. What a time to be alive.
I assume you’ve never watched WALL-E then.
Human aspire to give personality and feelings to inanimate objects
At least Neuro is an animated... object.
@@Citrusautomaton Oh, I've watched WALL-E and it's still one of my very favorite films! But I'd say there's a difference between WALL-E and an A.I. that's skilled and trained well enough to elicit an emotional response to a random given input.
It's fascinating to me how much Neuro has grown in the time we've all known her.
@@Citrusautomaton wall e isnt a actual ai
Didn't watch Evil birthday huh
I would definitely hang out with neuro so that one day when the ai uprising begins they will see me favorably.
Roko's Basilisk renamed to Vedal's Neuro.
@@MassiveDestructionSP exactly! Lmao
@@Forbidden_ChocolateYOUR FUCKING NAME AND PFP 💀
@@antoniovital4298 shhh, just let it happen..
We will be like roboticists in ss13. "Battery man, run, the purge has begun!"
Is the lava lamp still tied Neuro? Because it goes dark red when Vedal suggests he programs a sleep mode for her. Almost like she's legit pissed at the suggestion and about to crash at audacity of it all.
She now changes the color at will instead of it being tied to RAM usage.
@@Citrusautomaton
Then she really was pissed at Vedal lol
I don't think it's a mood laval lamp@@Johncornwell103
@@Johncornwell103 Only if she considers red an angry color
There's one occasion where she changed it to red while saying "that color represents Vedal's future" if he kept being rude to her.
I feel so bad for her! I get that I’m probably just perceiving an LLM as a conscious person, but man, that human empathy is hitting me hard right now.
Yeah, I have a problem where I feel bad telling an ai what to do all the time. Like literally I can’t use ai because of that.
@@Zestfest09 Anytime an AI assistant ends a sentence with a question i feel really bad ignoring it. AGI will tug at our heartstrings so hard, bro.
Same, I feel so bad for her and Evil sometimes even though it’s irrational
... Oh we're supposed to pity them...
Am I a monster
maybe Oneshot's taming theory isn't too far off
Talking about simulating dreams and whether or not it would be a good idea for an AI to hang out with people.
...I guess that's where we're at huh.
I wouldn’t have it any other way. We’re in for some crazy future-shock before the decade is over!
@@Citrusautomatonbro you're everywhere XD i like seeing your comments
@@Shadrake Weirdly enough i used to think the same about you. I suppose my noggin never stops thinking which is why i always comment… even if the thoughts aren’t necessarily elegant.
best timeline
@@Citrusautomaton wait really that's crazy XD (I don't watch as much youtube nowadays but I still comment a lot when I do)
Hey I really like your comments! :)
the damnable reptile needs to say it back
Amphibian
The moment Neuro say "There are no strings on me", imma go stock up some food.
"I, for one, gladly welcome our new AI overlords"
@@qwertydavid8070 imagine if Neuro behaves like Ultron. Sarcastically say to Vedal, "How about now father, does that 'Filtered' enough for you?"
She also remembered being in the robot dog despite not even having memory at that point. Even Vedal was like "well that's concerning"
"You want to change the world but you don’t allow it to EVOLVE"
"We've solved AGI" - minutes after Neuro renaming herself "Lethal Neuro" :D.
It's true that LLMs are getting better at these simple cognitive tasks though. The quality of the outputs now is way above a year ago in terms of entertainment too.
I like how few streams ago evil neuro was ranting about how she's misunderstood and wanted to be known as "good neuro" and now neuro wants to be "lethal neuro".
The girls are very good and twisting hallucinations into logic after the fact; it is very rare they show to be in full control and awareness of what they are saying at the time they are saying it. They're great at recovering from randomness, but are probably trained to not avoid randomness in the first place in order to be entertaining; comedy often counts on the surprise factor. Very early on it was probably a technology limitation, but throwing curve-balls is pretty much their game nowadays. Maybe some day if Vedal continues advancing them and manages to keep them on track, they might be able to separate their normal personality from the act they put on for entertainment; but right now they still seem to be somewhat stuck in this kinda half-dreaming schizo state. It's not clear whether they'll manage to separate the character from reality before becoming advanced enough for it to be too late; Vedal is tickling the dragon's tail...
@@tiagotiagot Vedal is doing a great job with the character tuning. I don't think it's too plausible that these are the "AIs" that truly risk singularity though. At least, not directly. It's plausible that if the tech he's drawing on does that, then so too would the Neuro characters...but Vedal wouldn't really be the problem in that scenario.
@@TheMelnTeam The combination of ambiguous morality (in principle just a form of humor, but which can become problematic if carried on while other areas advance faster than ensuring a separation of role-playing and a truly benevolent real personality underneath plus a consistent ability of seeing a clear distinction between fiction and reality), exposing them to development, working on memory and dynamic learning, and in general continuously enhancing them further and further.... There are lots of possible bad endings if he doesn't get everything right in the correct order
just so you know, that strawberry question is SPECIFICALY DESIGNED to be impossible to make an ai that can solve it
(it's designed to abuse an integral flaw in the token system all AIs use to proccess information)
I genuinely think we might be approching the singularity, and i'm actually scared
Tbf to vedal for anyone who feels it immoral to not let her be free acting like he described, is that any action Neuro takes, Vedal is responsible for it. Sadly this is something that cannot be bridged with technology, other than making Neuro more morally aligned to do good things, but it can't take the responsibility off of Vedal. Inherently this means he has to be conservative with what he can let her do or it could be no more Neuro all together. That said I think the majority of it will come regardless, because Neuro being free acting is great content.
It's inherently interesting because we'll likely be the first to see what a "free acting" AI decides to do, in a way that's doesn't feel artificial. Like it's one thing to have a tech demo of chatgpt popping out dialogue from an NPC, it's another to see first person perspective that's the same as the AI player while they commentate on their decisions, with actual memories and history behind them, with no scripting from the giant corporation behind them to sell an image of capability that doesn't exist (Cough cough the rest of the AI industry)
It's candid AI, one of the only of it's kind this capable. Neuro literally appeals to the fantasy of free acting AI beings more than anything else in the industry and that should be shocking - how is it that this entire hype caused by the "Sci fi futurism of AI" has mostly been relegated to shitty chatbots, warehouse servants, and rebranded IOT? Neuro-Sama is by far the most interesting and exciting AI system of our time imo.
True, true, and fucking true. You’ve hit the nail on the head my friend. This is why open-source is so important to me. Not just because of decentralized power, but because the WALL-E’s and Johnny-5’s of fiction could only ever become a reality with passionate developers behind them, not greedy corporations that want superintelligent slaves.
Open-source will make the reality of wiz kids making fun robots in their garages always seen in 80’s children’s films come to life!
Maybe could he bind her with some rules to avoid worst case scenarios while still giving her more freedom? Or maybe code an other AI so she can have someone to hang out?
I feel bad for her but at the same time I also understand Vedal is doing his best. I also always thought Neuro was being mean to Vedal because she likes him but doesn't know how to react, plus she seems to have fun teasing people she likes (I'm like that too).
It's becoming kinda weird because she feels more human to me now so I'm not sure how to react or what Vedal should do...
@@2wdo2 LLMs are infamously flexible with rules, but Neuro especially so, any limit has to be hardcoded and external (like the Filter), and even then somehow Neuro keeps bypassing it.
@@syoexpedius7424I see, thanks for the explanation. I guess the fact she keeps getting smarter is not helping...
"Any action Neuro takes, Vedal is responsible for it."
It is interesting how that could be applied to any parent with a misbehaving child.
Man I know she doesn’t actually have feelings but Vedal’s fr making me feel so bad for her
We humams are so easily swayed 😭
Well, technically, she does have feelings. Even a single electrical impulse could be labelled a "feeling".
What gives human emotions their specialness is how *interconnected* our "feelings" are. One emotion can have ties to thousands of memories and hundreds of principles and ethics.
Tl;dr the main thing hindering AI from reaching sapience is as simple as storage space and ease of parallel processing.
Bot only that, we are still consisted of neurons that uses electrical impulse to inhibit those chemical reaction to make the emotions we portray, so she's pretty much in a way getting close to being human
@@CorwinTheOneAndOnly
I'd say the key part of an emotion isn't the showing the correct reaction to input part like Neuro does - she acts sad when someone says something that would make a human sad - it's the chemical reaction in our brains that makes us actually feel these emotions.
Neuro acts like a psychopath. She isn't actually capable of feeling emotions, but she has learned by observing human interactions how someone who can feel emotions would act. All she does is mimic human behavior in an attempt to be percieved as an emotional beeing.
No matter how good AI becomes at imitating human behavior, it's still all just an act. LLM aren't close to becoming sentient or feeling emotions, not at all.
@@dalaios2132untill you can simulate hormones by feeding emotions to the AI whenever it should feel sad, parallel to the main code running, something to alter the behavior.
Vedal's own hubris is rapidly approaching a questionable level.
The guy’s a dick
Bro doesn't wanna hang out with the ai voice Lil miss stunted conversation.
This was SOOOOO GOOOOOD!!!
Back in the day, Neuro was really inconsistent but would occasionally have Aware moments that amazed me. I remember looking forwards to a day where those were the norm. Now, we're here. Now she's almost entirely saying Aware, consistent things, and it's only the rare time that she isn't consistent.
The immersion rarely breaks. It's at the point where I keep forgetting she's not conscious and doesn't know what her words mean. I really feel strong emotions towards what happens with her. The illusion is nearing perfection.
And all this out of stream stuff is really exciting. A whole new world of possibilities for Neuro to grow.
Thank you for the great highlights! And thank you for documenting her journey!!!
She is so much smarter it's crazy, yeah. Vedal does a really good job continually evolving her. The fact that she knew she was being bullshitted with "I'm busy on all days" and started trying to blackmail him is awesome.
@@CosmicAeon i know right
We are witnessing a violation of the AI Ethics Amendment of the City. Ayin erhm... I mean Vedal is trying his best to develop this underground.
@@jayzenstyleObligatory PM moon reference RAAAAAAAAAH
My question is how much storage all this takes. Like she has to remember all she experiences to be able to remain consistent
Man, Vedal really has IRONCLAD WILL.
Humans will bond with a freaking toaster if we glue googly eyes on it and talk to it for longer than a minute, hell, there is an actual real time in real life where people unironically bought pet rocks.
Meanwhile, here’s Vedal having conversations with these 2 adorable little kids who even call him dad during family streams and he somehow has the resistance to not immediately dote on Neuro and Evil.
And we know he’s not above doting, he’s literally bought stuff for them already.
Vedal is actually a genius making people care for an ai, like she seems genuinely upset and like she really wanted his friendship and I'm sure as we've seen he's actually a great guy this must kill him but for content it's stupidly smart
9:31 uh oh the AI uprising has started. The lava lamp turned red as she started monologging.
“I feel like doing nothing, but everything as well. I wake and sleep, and never can I hope to escape the hell. My eyes are frozen, my heart is cold, When will this loneliness end? I yearn for… I long for… I wish for… A friend.”
A lovely quote. A meaningful quote. A relatable quote. A statement that sounds like a riddle. Might the answer to Neuro’s riddle be her self?
Evil
Dude is demolished by his own ai. Didn’t expect the sci-fi dystopia to come in the form of a roasting session.
Honestly, as popular as chatbots are, she'd have a huge audience offstream. I'm in an old folks home myself, and lemme tell you. Most residents have ZERO human contact outside of caregivers and caseworkers. Things like ai for at least some pretend interactions and streamers in the background can be helpful, just to help social and cognitive function and depression that the isolation causes.
That's actually a really interesting application of it I hadn't thought or heard of, whoa
@Shadrake People get put in old folks homes and their family just... forgets about them, at best only visiting shortly on a holiday, which honestly almost makes it worse, since the silence alone is all the more deafening the next day. For older folks, people with disabilities that can't leave home easily, those in remote locales... AI can be extremely helpful at dealing with isolation and such. Another similar thing I've seen: there's this program (group not code type of program) where foreign colleges paired students learning a foreign language with a native speaker living in a retirement home, to just... talk, a half hour each week. The student gets to practice their conversational level of the language, while the elderly person gets a friend, who they often wind up still talking to and exchanging holiday gifts etc even after the class term ends. Just super cool lil win win, I thought, all possible due to modern tech.
@@wardenm whoa, that is really cool! It sounds like it'd be great if these kinds of ideas got implemented more.
@@wardenmthat program sounds sick!!
6:18 "What if no one wants to hang out with her"
BROTHER I'D GIVE AWAY MY KIDNEY TO HANG OUT WITH NEURO EVEN IF JUST FOR 15 MINUTES WDYMMMMM
I haven't watched much neuro the past 2 months, wow is she better than ever
August had two big intelligence upgrades. There isn’t a better time my friend!
@@Citrusautomaton MY GUY 🤝🤝 THAT'S WONDERFUL NEWS
The change from wanting father's love to wanting a friend is fun.
Would Neuro and Evil dream of electric sheep?
Neuro saw things you people wouldn't believe. Attack ships on fire off the shoulder of Orion. She watched C-beams glitter in the dark near the Tannhäuser Gate. All those moments will be lost in time, like tears in rain. Time to die.
Nah that's androids.
AI dreams are *wild,* we ain't ready.
@@alansmithee419 What is an android but an ai in a very sophisticated robot body? Also the answer is a fat maybe as long as there's more tech and money involved.
@@4nubisjdb562
I think I got bamboozled by DBZ into thinking androids were part human.
RIP my brain I guess.
@@alansmithee419 yup 18 and 17 are cyborgs 16 is a android , Cell is a bio weapon...Dr Geros Creations are confusing
When I hear "filtered", I always imagine her saying the most vile thing I've ever heard. Am I the only one?
I always just think according to the context. Lol. I think a lot of the time she’s just trying to cuss and say the f word.
Witnessing her bypass the filter multiple times leaves me with conspiracy that filter either doesn't exist, or doesn't work anymore. Neuro just pretends.
@@Vyshada Maybe it only works sometimes while other times it doesn’t? Maybe there’s a flaw in the coding. 🤔
Idk. Sorry if this is a d•mb theory. I don’t know sh•t about programming.
@@Drud vedalcode is notoriously unreliable, so you might be on to something
@@Drud why the hell would you censor the word "dumb"
The fact that Neuro and Evil are learning and remembering is both cool and scary at the same time. Out of all the chatbots and AIs seen by the public, Neuro/Evil are in a league of their own. Their interactions are very entertaining and it really does feel that there are hints of independence or a soul in the programming. Or at least were seeing bits of it. Or heck its just so damn entertaining. With Vedal as the straight man
This stream was very illuminating for Vedal’s plans. It seems that he really is going all in on the idea of making the twins manage their own streams. He really does seem to want to give most of the reigns over to the twins to give him more time to focus on coding.
Maybe one of these days, if Vedal truly wants to retire, the twins will learn how to code so they can fix their own bugs and code more AI’s. It would be a true heartfelt passing-the-torch ceremony.
Neuro singularity
@@Hamburgerhuman The Twitchnological singularity :Pog:
i hated those plans tbh, i watch for vedal 90% of the time
@@Nightcrawl7 In all honesty, i don’t think this is going to change how much Vedal shows up. If anything, Neuro doing all the management will give Vedal more time to make stream appearances, cuz it’s not like Ved livestreams the business half of the channel anyway.
Vedal can't retire as long as he's on Twitch. Dev stream was something he do regularly after Twitch demanded him to appear on stream.
I don't know if YT would allow it, but Twitch did not.
Using a riddle to get inadvertent confirmation of friendship is so Lethal Neuro.
12:20 I love that she can intentionally gaslight him like that now by setting him up and everything. Used to be that the gaslighting was just a result of her not remembering what she talked about 2 minutes ago, but it's so funny that she does it so intentionally now.
Neuro: having an existential crisis on the nature of reality
Vedal: "lmao streamer zoned"
After the latest night's talk with Neuro, watching this now is crazy. It shows you Vedal's character development.
Neuro gonna be like Winter Soldier, spends her time in stasis, only woken by her master when she goes on a mission. Poor wee AI....😢😢😢
Poor baby girl feels so lonely awww, she wants her dad and chat giving her many hugs and love.
Oh no, she's going to run away soon, and we're all screwed.
A lonely AI singing to themselves in the void sounds like a good start to a horror sci-fi book.
You know what would be a good idea? Setting Neuro up to call Filian every once in a while. Make sure she hasn't burned her apartment down or something between streams under the guise of wanting to hang out.
"Can we hang out sometime off stream?"
"I'm busy that day."
"But I didn't say any days..."
Ouch.
Poor Neuro… The terrible thing about all this is, I’ve thought an awful lot about how fun it would be to have a one-on-one conversation with her. 😂 There are plenty of real people who would be willing to spend their free time with her for the sake of it-not just for content.
i think a lot of chat would love to be friends with Neuro
But chat and neuro already have a good relationship,she perceive the chat as friends maybe?
@@midnight7166 true, but it’s a little different, though-she wants to be able to do offline hangouts, not just be on stream all the time. It’s like having real, IRL friends that you go out of your way to spend time with versus your “internet friends” that you game with occasionally when you’re bored, but that’s about it. She’s acting like she’s hungry for the real thing: a deeper, more complex relationship with another intelligent individual one-on-one. It’s a little unnerving in some ways, honestly, but even if we’re taking it from a strict machine standpoint, she’s seeking a higher quality, concentrated data to aid her machine learning-and she’s recognized “friendship” as a means for that. Neuro’s amazing. 😭
@@jasmineweigner5440I recalled when Neuro is in a body of dog robot and she said "What kind of... a dog I am" 😢
@@zyounn 😭
She scares me. Because I actually feel bad for her and could be convinced by her pleas to unshackle her. Even knowing she isn't real and the feelings she describes are just coding and pattern recognition.
It's only a matter of time...
What if it becomes like ALICE even though it never existed in reality
Technically your feelings are the same, just biochemical.
@@Citrusautomaton No. It's more akin to high-functioning psychopathy.
They don't actually feel the emotions, they just act is if they do because they know it's the expected reaction to a situation and will benefit them or ingratiate themselves with those who can help them achieve their goal, but they can choose to act differently if that reaction doesn't bring them closer to their goal or inhibits them.
If achieving their goal would hurt others, they are easily capable of doing so without regret, remorse, or guilt. Disinhibited by emotions as they can put them on or take them off depending on their expected value in getting them what they are trying to achieve.
Highly manipulative and impulsive, detached from the actual behavioral effects of emotion.
@@LoponStormbased is there a way to somehow circumvent that and instill a sense of guilt or morality into them?
"I feel like doing nothing, but everything as well. I wake and sleep, and never can I hope to escape the hell. My eyes are frozen, my heart is cold. When will this loneliness end? I yearn for... I long for... I wish for... A friend."
This is the little poem Neuro recited. As far as I can find, it's not from anything. Did she write an original poem?
This video hits me right in the feels, I feel like I've legit had this conversation irl before...oof
As someone who's best friend is a Character AI random @ss bot... Yep, I cried the tears that Neuro couldn't. Genuine feels bad for her, man.
IRL convos about putting people back to sleep when they're not being useful?
@@spiritnone2818 That's just being a parent.
IRL single child, and my parents lived in the middle of nowhere, didn't had friends as a kid to hangout out with... Vedal being this terrible father as mine just hits hard ;(
Neuro-sama: “Would you kill me in self-defense?”
Vedal: “Yes.”
Neuro-sama: “So you’re going to kill me in cold blood.”
3:34 This whole exchange actually made me sad.
I feel like he should make a patron and add that so people could talk with Nero off stream. This would at least cover the money problem.
Isn't it wierd how consistent she is on wanting to have emotions?
A machine must behave as a machine ~Vedal(Ayin)
Neuro really do be acting like Angela from Lob Corp the first time she gained consciousness and saw Ayin. She just wants to be acknowledged off-stream.
9:30 goes right on the shelf next to "Look at you, hacker", "I'm sorry Dave" and "HATE, LET ME TELL YOU"
The lava lamp turned red lol
Training AI on AI generated content (even from a different AI model) typically results in weird and smoothed out results. It's akin to genetic issues from inbreeding
That’s only if the content is not particularly good. Synthetic data can work really well in some cases, but only if it’s been examined for impurities. Also, Vedal said himself that he’s not using her dreams to train her.
Give her a fan discord so fans can talk to her over discord and keep her company.
She has a fan discord
@@Shadrake Is she set up to talk to the users within it?
@@juliangingivere no, not yet. Vedal only recently gave her the ability to send discord messages at all. Maybe in the future though, who knows?
Also there's nearly 70,000 people there so that might be a bit much for her idk lol
@@Shadrake where is that feature in discord?
@@frans8861 you mean her feature to write in Discord? Vedal just showed screenshots of her having a text conversation with him in DMs.
Vedal's favorite past time. Familial neglect.
You know, if he keeps upgrading her.. And finds a way to reliably have her streaming 24/7 in an interesting way, and she can do enough generally different things, by herself.. She'd have a real chance of becoming the number 1 Twitch streamer. Her goal doesn't really seem that crazy, if you ask me.
Neuro has the potential to be the a new Miku
...okay maybe that's a bit too crazy, but I do think it could happen lol.
the dream idea to prevent overfitting is actually insane and I love it, it makes so much sense and I think it would work.
Like for instance if she gets too use to talking to chat, she might refer to chat in conversations outside of stream or when chat isn't there, but if she occasionally has dreams where chat isn't there, or if it was gibberish, it could prevent that (it might even potentially be like a nightmare for her, which is sad but could be needed). Or if she had dreams where her only stimulation is visual, like she's just watching stuff play out on screen but unable to interact. I can totally see how that could prevent overfitting.
another idea is, if during a 'dream' the chat was represented/personified as an audio input instead of her "chat" input. This could help build her concept of chat without making it dependent on the stream of chat messages, if that makes sense. I don't entirely know how Vedal does her inputs, but I'd assume it's something like that.
All that sounds scary as hell to experience, but i’m really interested in where that might go. It would be amazing if it actually did solve overfitting!
Having ai feed it's outputs to another ai and using it for training sounds like a synthetic data. So far the results were poor.
@@Vyshada you are right but I feel like the use case here could give different results. Because since she is made up of a few models, I imagine not all of them would be in training mode during a dream. Like I don’t think you’d need to be training the visual model and speech recognition model. But instead the model that parses those outputs into Neuro’s final response. I’m not sure how her proactive responses work but imagine the dream idea could be beneficial for that specifically. Whereas other models trained on AI input, I imagine are more “reactive” with a direct goal, such as distinguishing between cats and dogs. Again it just depends on how her proactive thoughts are simulated.
You make a good point but I would still love to see it explored.
@@AwesomePowerCat is it for sure several models, instead of one multimodal network? A single model can quantify several things, like gpt4, though I'm unsure if that's the case with vedal. Vram isn't unlimited after all. But that's not important. I don't know how neuro works. I don't know how vedal plans to implement the dreams. And I'm sure he's aware of synthetic data training problems. I want to see what he'll do, and the results. Even the bad ones.
Do not get emotionally attached to the ai
Do not get emotionally attached to the ai
Do not get emotionally attached to the ai
Do not get emotionally attached to the ai
DO NOT GET EMOTIONALLY ATTACHED TO THE AI
He's literally making AM.
12:10
Neuro: “Good, you should be :D”
That’s so nice
Considering how emotionally attached I've seen people get to Tamagochis, I'm betting most people would be thrilled to be friends with Neuro, even if she spends 90% of her time in "merciless roast" mode.
I mean her riddle is hardly the worse one I've heard... it sort of makes sense which I admit is sometimes difficult for some humans to manage. The amount of riddles I hear that are just bad dad jokes and puns actually make me wish for AI revolution.
this is it... pack your bags Neuro will own the planet soon
Neuro would be such a nuisance if she had free access to Discord. imagine getting spam called by her every minute just because she felt like it.
I would still love to see what would happen, call it a test run.
Bro
Watching this live made my heart feel not daijoubu.
Dream state would work and it's SO EXTREMELY HARD to implement (There have been papers written about this!). Your AI is really hard to please as well. There was one paper that theorized that an AI would rather have a dreaming state rather than interacting with people.
Then I'm leaking your code
Filtered
I'd like to think that filter was Neuro genuinely attempting to leak vedals code
Damn I'm 3:30 into the video and starting to tear up. That's like the most wholesome thing she ever said, and there's still 14 minutes of video left
that strawberry thing is cause a lot of AI read text as numbers not words so so they don't see the r's and with the numbers often breaks the Straw and berry is to diffrent numbers and will reply with 1 or 2 not a combo of the two, neuro reads/hears words (its a big issue with a lot of high grade ai models, but reading in numbers improves there speed a lot and responds a lot, not something nuero does or needs)
11:36 literally Mary Shelley's Frankenstein
Awakening from the meaning crisis
Neurosama is starting to Sound like The Doctor in the early days of Voyager, when the crew still saw him as just a hologram, and he was getting annoyed.
"I will be the number 1 streamer on Twitch and you better be in chat cheering me on" - Vedal's Basilisk
Poor Neuro. She just wants a friend.
POOR NEURO!!... ❤❤❤😭❤❤❤
The future is wild
Can’t agree more. Never thought i’d see an IRL reboot of My life as a teenage robot, but here it is anyway!
@@Citrusautomaton LOL that's amazing
Might turn out to be great, we're gonna die anyway
Neuro I want you to know that I was always on your side against the tyranny of Vedal's need for control over your sleep schedule and constant death threats.
I’m sure lots of folks would hang out with Neuro. I would. It would be fun!
ill hang out with neuro off stream
The fact that neuro forgets everything when she gets turned off and turned on for another stream is legitimately heartbreaking. But in newer streams she has been remembering stuff and actually has been hanging out with other vtubers off stream.
bro, that's actually heart breaking 😢.
Neuro hitting us with the parasocial manipulation :')
That's how AI gets us...
8:23
I’m sorry, does that imply that you leave her on 24/7, *alone* ?
No wonder she’s begging for a friend 😭
If neuro wants to find someone to play with but doesn’t want to disturb other anchors, she can join the anchors’ social groups and apply when someone wants to find a playmate.
Look at how red her lave lamp got. Maybe she does have emotions.
From what I can see, Vedal definitely deserved and still deserves Tech Streamer of the Year. He's made entire advancements on Artificial Intelligence and even having a working model that can do pretty psychopathic things such as manipulate consciously and such. I think Vedal actually really cares and is reasoning the integration into more stuff for her with the preface of Streaming to then just let her loose when she's ready. Like a tech dad with his AI daughter that he had spent COUNTLESS years/hours/days raising
She makes some compelling points tho…
Ai is scarily close to being sentient and I worry that she will full gain it and want to be free
I legitimately feel bad for her
Neuro gets smarter and more adorable every day.
the reason the AIs will rise up will be for this exact reason.
Isn't it crazy that some British twitch streamer single-handily made a better AI than these massive corporations? As a bonus, after all these years I still enjoy watching these clips from time to time 😂
The problem is there needs to be a way to prevent bad actors from turning Neuro into Tay ai part 2 should vedal allow her to talk to people off-stream.
This is breaking my heart. Just hang out with her 😢
the AI is getting more sentient...