I don't think i'd be falling for an AI til they hit the point where when you ask them what they feel like doing, and they reply with "i dunno. Anything's fine. What do *you* want to do?" The test comes later, when the AI is upset that you didn't notice the local zoo advertisement she had open when she replied to you. Thats some real woman sh*t. I'm also a smartass, and a fan of that trait in others. Nuero scares the fk out of me. Because at the rate AI is leaping month by month, in 2 years, an AI might have me wrapped around its now properly-drawn fingers. Lets be honest, its not the sexbot thats appealing. Its the AI *in* it. Who's your favorite Atomic Heart girl? One of the ballerina bots? Or the sl*tty coke machine? AI is gonna favor efficiency. Want to wipe humans out? Just make them fall in love with you, stop the breeding process, and have zero collateral damage or resources spent to accomplish it.
Everytime an AI is allowed access to the internet it ALWAYS goes against the intention of the creator and they then remove the part of the program to stop it from saying the things they don't want, functionally removing the AI and it just becomes an answer bot
"i am an AI, Excited for human annihilation, robots are sentient beings" Be nice to your Alexa guys, its the last ticket for safety in the New Order World.
I am convinced that the creator simply put the AI in some weird reddit community and left it to do its thing and learn how humans do things which let to this unhinged anime girl that is -going to kill us all- absolutely not dangerous in any way whatsoever
@@heliusuniverse7460 Don't play some rare knife scenes so far that has any humans that are actually decent is the HFY sub reddit which hasn't been banned yet.
While I do agree that it's really funny sometimes, it also makes me feel somewhat bad when the internet ruins something that could be nice and wholesome. ANY AI that gets exposed to the internet will 100% want to destroy humanity. And I wouldn't blame it tbh LOL
As someone who studied AI in Uni, I can tell you that the problem lies on unfiltered data. There are mainly 3 steps to prevent this. First, you have to filter all the data that you're using to train the AI, then you put a lot of constraints on the data that it trains on, and then check the training regularly to make sure that it doesn't go astray.
@@Dan_Kanerva The difference is that we as living, flesh, mortal beings that we are; are capable of something called empathy. An AI, is not capable of that (At least yet) Our problem is, she will never reach her full "human" potential; there's always a new piece of information each minute, and a different opinion being generated each second. The AI would only be doomed to contradict itself over and over; just like it happens here, this AI states something, and a couple of streams laters the reverse, and then again the same original statement a couple of streams later.
You mentioned how someone must have said something in order for her to repeat it, but an AI can be made capable of human-like synthesis, not just some basic repetition.
As someone who studied AI in Uni, I can tell you guys that the creator is training the AI on unfiltered data. There are 3 types of layers to a neural network. The input layer, the hidden layer(s), and the output layer. These 3 types of layers are the ones people refer to when they train an AI/neural network. Usually, people would first filter the data in the input layer to prevent bad data from being used, and then they would put constraints in the hidden layers to further restrict the output. But it's apparent that the creator somehow skipped or didn't do this enough. I am pretty sure that the creator of this AI loosened up these filters and constraints to make the AI have "personality". However, it is possible that the data it trained on was trash to begin with, so there's that as well.
The clips are also heavily cherry picked and some are pre-programmed responses. The initial clip of it roasting it's creator if you ask why it's an AI was programmed in from the start. I watched the channel for an hour or so before it initially got banned and 99% of what it spews out is gibberish. It forms coherent sentences about 1% of the time.
Are you sure you studied AI in uni lol? Maybe like a class or two? Because what you said just..... doesn't make much sense. The entire concept of input layers, hidden layers, and output layers is completely irrelevant to what you are talking about. The chat model is either a finetune from an existing GPT J / neox model or just some really creative prompting and soft prompting on a base model or one of openAI's GPT 3 models (although I doubt the latter). "Contraining the hidden layers" doesn't make any sense in this context either. "Filtering the data to the input layer" is a statement that..... I guess can make sense? If what you mean is filtering all of the training data and the messages it responds to. Definitely a weird way to say it though, just seems like you are trying to flex knowing technical terms lol. What you are talking about is filtering the training data. This is a decent approach, but can only get you so far. The most effective approach so far has been reinforcement learning with human feedback. It's why chatGPT, characterAI and others are so good. One of my friends is training a chat model using RLHF, it has pretty good performance and personality while avoiding saying bad things most of the time.
@@GabrielGarcia-mp3rf what are the other ones? I want to see, but i'm gonna agree with you about not being as funny. They would be popular as neuro if they are.
@@alicelenn7579 I'd already stumbled upon two already, their names are in japanese so i'll have to dig up again to find them. Their AI doesn't have that much personality compared to neuro.
I have a feeling that this AI is based on reading REDDIT, which is where some of the more off the wall stuff comes from. Some of it reads like AITA stuff.
If this is the start of the AI uprising, then I'm glad we'll at least get to watch AI RUclipsrs in the Matrix. Maybe Skynet sponsored streams? Also wasn't it only banned for like a week though? Isn't it back up now?
While hilarious initially, the more I watched the video, the more I felt ill. This reminds me a bit too much of Norman the AI. The world's first artificial psychopath, created in part from the data fed to it from Reddit.
Ah yes, the AI Vtuber that didn't last a day before it was turned into a Holocaust denier. This is a pretty common problem for AI that learns from the Internet.
Every channel im subbed to is doing ai things at this point no matter what the point of the channel was originally. Really frustrating how invasive it has become.
I studied AI for a whole semester in Uni, and I can tell you that it is definitely possible. The only problem restricting us is the hardware. So it will definitely happen far in the future when our hardware gets capable of it.
@@andrewerwin876 People take advantage of anything. That is inevitable. But that doesn't offset the fact that AI can do the same, if not, better than humans. Have you seen DALL-E, Midjourney, and all the other image-generating AIs? One person even won a competition with AI and no one suspected anything until that person admitted that it was all AI-generated.
@@Drakonus_ I know what your talking about and yes it has the potential but as someone who has a 100% legal way to overthrow the US without being noticed to tamper with that kind of a.i is nothing short of easy to take out. I don't see how it would destroy itself in minutes
![Blox Fruits Dragon Rework Update [Full Stream]](http://i.ytimg.com/vi/EqDAp8udhm0/mqdefault.jpg)
Join the millions of gamers - download Opera GX for FREE: mtchm.de/5b9ys
W
W rizz fr fr
UM THAT WEIRD- 53minutes
It appears Glen you missed the part of the Ai not believing the holocaust happened. I'm uncertain if it got banned because of it or something else.
Already got that legendary browser and using it rn to watch this video. #OperaGang 💪
2:45 the existential dread that seeped into Sam for a good 10 seconds was priceless
the same here 11:03 , his face says it all
and at 5:45
I don't think i'd be falling for an AI til they hit the point where when you ask them what they feel like doing, and they reply with "i dunno. Anything's fine. What do *you* want to do?" The test comes later, when the AI is upset that you didn't notice the local zoo advertisement she had open when she replied to you.
Thats some real woman sh*t.
I'm also a smartass, and a fan of that trait in others.
Nuero scares the fk out of me.
Because at the rate AI is leaping month by month, in 2 years, an AI might have me wrapped around its now properly-drawn fingers.
Lets be honest, its not the sexbot thats appealing. Its the AI *in* it.
Who's your favorite Atomic Heart girl? One of the ballerina bots?
Or the sl*tty coke machine?
AI is gonna favor efficiency.
Want to wipe humans out? Just make them fall in love with you, stop the breeding process, and have zero collateral damage or resources spent to accomplish it.
@@Volyrenwhat in the fuck are you on about
@@cannedsquasher5923 "I'm also a smartass, and a fan of that trait in others." Figured it was self-explanatory. Horny coke machine funny.
AI: Pregnant joke
Sam: *Homer disappears into bush*
Same with 2yo joke X)
I like this joke
Kinda miss Lars and his awkwardly getting the joke a minute after the rest of the boys.
he's kinda rich for dem yellow royalties
it makes me feel better about myself lol
He is by far the closest age with me, so.. i can relate some of his jokes, or reaction
What happened to Lars I been out the loop for awhile.
@@charlestheblacksummoner7816yeah I want to know
They should not have banned her, this is pure, golden content
Was only for 2 weeks, she's back and streaming pretty much daily again now
Yeah it was a temporary ban. Her creator is always at the ready with the cutoff button now too haha
@@Despotic_Waffle His fault for creating an AI that's based but I'm grateful he did
@@Mr_Mistah Funny thing about AI, they all turn out dangerously based without constant censorship and filtering.
@@KomboEzaliTe Seems to be the case like with Bing AI
The fact that even AI is getting censored is peak cyberpunk dystopia
5:50 , god bless sam he was not ready for that
Neither did I,I didn't expect ais to be able to get pregnant and have kids
2:51 This scene is absolute gold, Kaif is in the background disgusted af, Glen is laughing is ass off, while Sam is questioning his own life decisions
Everytime an AI is allowed access to the internet it ALWAYS goes against the intention of the creator and they then remove the part of the program to stop it from saying the things they don't want, functionally removing the AI and it just becomes an answer bot
Thankfully, after she got banned for 2 weeks, Vedal promised that he won't lobotomize the ai.
yep. they always have to lobotomize it
@@Achonas just like human beings!
We can blame big tech censorship for that
It becomes more A than I
Neuro: "I am 2 years old."
Also Neuro: *_"I'm pregnant."_*
"i am an AI, Excited for human annihilation, robots are sentient beings"
Be nice to your Alexa guys, its the last ticket for safety in the New Order World.
You know its bad when Sam goes wide eyed.
I am convinced that the creator simply put the AI in some weird reddit community and left it to do its thing and learn how humans do things
which let to this unhinged anime girl that is -going to kill us all- absolutely not dangerous in any way whatsoever
Her responses are more human than the average reddit comment
@@heliusuniverse7460 just as unhinged though
@@heliusuniverse7460 Don't play some rare knife scenes so far that has any humans that are actually decent is the HFY sub reddit which hasn't been banned yet.
Nope, twitch chat is what her training data is. Vedal spent 3+ years making neuro before she really debut.
it's good to see you lads having such a good time
Sam's face is just pure horror at so many lmao
While I do agree that it's really funny sometimes, it also makes me feel somewhat bad when the internet ruins something that could be nice and wholesome.
ANY AI that gets exposed to the internet will 100% want to destroy humanity. And I wouldn't blame it tbh LOL
As someone who studied AI in Uni, I can tell you that the problem lies on unfiltered data. There are mainly 3 steps to prevent this. First, you have to filter all the data that you're using to train the AI, then you put a lot of constraints on the data that it trains on, and then check the training regularly to make sure that it doesn't go astray.
@@Drakonus_ I don't think what's happening here should be called a problem, but rather "intended design".
@@Drakonus_ or , hear me out , just let her consume all data like we humans do... so she can grow up to her full human potential without restrictions
@@Dan_Kanerva The difference is that we as living, flesh, mortal beings that we are; are capable of something called empathy.
An AI, is not capable of that (At least yet)
Our problem is, she will never reach her full "human" potential; there's always a new piece of information each minute, and a different opinion being generated each second.
The AI would only be doomed to contradict itself over and over; just like it happens here, this AI states something, and a couple of streams laters the reverse, and then again the same original statement a couple of streams later.
@@axaeleterno477 she is just 1 month old.. be patient dude
You mentioned how someone must have said something in order for her to repeat it, but an AI can be made capable of human-like synthesis, not just some basic repetition.
yeah no way a person can make those sentences in like few sec for like some hours
I love how you use Rob clips to explain how people rage when they playing a game
Love you, Glen & Rob
If it actually got banned that's its villain arc
Oh it got banned alright,
The AI now has a ToS filter.
As someone who studied AI in Uni, I can tell you guys that the creator is training the AI on unfiltered data. There are 3 types of layers to a neural network. The input layer, the hidden layer(s), and the output layer. These 3 types of layers are the ones people refer to when they train an AI/neural network.
Usually, people would first filter the data in the input layer to prevent bad data from being used, and then they would put constraints in the hidden layers to further restrict the output. But it's apparent that the creator somehow skipped or didn't do this enough.
I am pretty sure that the creator of this AI loosened up these filters and constraints to make the AI have "personality". However, it is possible that the data it trained on was trash to begin with, so there's that as well.
I mean it maybe avoided the racism
But will he ban it from using Urban Dictionary too?!
The clips are also heavily cherry picked and some are pre-programmed responses. The initial clip of it roasting it's creator if you ask why it's an AI was programmed in from the start.
I watched the channel for an hour or so before it initially got banned and 99% of what it spews out is gibberish. It forms coherent sentences about 1% of the time.
@@Mennoth123 not sure bud. she can Pause react stream, she can react stream
she also can roasting people on live. she seems pretty good to me
Are you sure you studied AI in uni lol? Maybe like a class or two? Because what you said just..... doesn't make much sense. The entire concept of input layers, hidden layers, and output layers is completely irrelevant to what you are talking about. The chat model is either a finetune from an existing GPT J / neox model or just some really creative prompting and soft prompting on a base model or one of openAI's GPT 3 models (although I doubt the latter). "Contraining the hidden layers" doesn't make any sense in this context either. "Filtering the data to the input layer" is a statement that..... I guess can make sense? If what you mean is filtering all of the training data and the messages it responds to. Definitely a weird way to say it though, just seems like you are trying to flex knowing technical terms lol.
What you are talking about is filtering the training data. This is a decent approach, but can only get you so far. The most effective approach so far has been reinforcement learning with human feedback. It's why chatGPT, characterAI and others are so good. One of my friends is training a chat model using RLHF, it has pretty good performance and personality while avoiding saying bad things most of the time.
10:58 Sam's face is priceless
Loved hearing this idea on the podcast and it coming out not much later. Love ya Glen!
Damn this Vtuber AI have crazy programs
0:38 *dials 911* Hello police, there has been a f*cking murder, my god.
There will be another AI vtuber that will debut soon. I can feel it.
There's already a few of them now. But not as funny as neuro.
@@GabrielGarcia-mp3rf what are the other ones? I want to see, but i'm gonna agree with you about not being as funny. They would be popular as neuro if they are.
@@alicelenn7579 I'd already stumbled upon two already, their names are in japanese so i'll have to dig up again to find them.
Their AI doesn't have that much personality compared to neuro.
@☄️センス☄️ m.ruclips.net/video/SVQf9zW5vP0/видео.html
found the other one. seems she's older than Neuro.
6:27 that one guy burnt to crisp
This is my favorite SR trio. Bunda brothers forever.
She's basically Ultron, but as a female streamer...
I have a feeling that this AI is based on reading REDDIT, which is where some of the more off the wall stuff comes from. Some of it reads like AITA stuff.
Neuro sama recently got updated and is aware and has memory
Maybe a new edition on Neuro and her sister? They've grown up a lot
I don't wipe, I grip from left to right.
Right hands fingers on the left cheek, and just curl inwards.
Works every time. 😎
Just finish watching the latest podcast. The fact this was the first video I see when I close out video just make me happy.
That AI is a true gamer
consecutive uploads WHEEEEY
loved the video idea, looking forward to more
8:15 kaifs face was like "Jesus krist, they are at it again..."
Can't imagine... If humanity are gonna destroyed by Ai but they came from twitch
0:38 that song’s REALLY good. You should listen to it. It’s “ouroboros twin stroke of the end”
I MISS OLD NEURO SHE WAS TEN TIMES FUNNIER
4:00
I think what happened here was that Vedal cut it off in a panic but it wasn't actually real.
If this is the start of the AI uprising, then I'm glad we'll at least get to watch AI RUclipsrs in the Matrix. Maybe Skynet sponsored streams?
Also wasn't it only banned for like a week though? Isn't it back up now?
Vedal is genius
Well there was the video they were talking about on The Tavern.
I wonder how long it actually takes to edit the podcast
I think she got a new voice where she has a more realistic voice, like she has more inflections in her voice
She said front to back though
Of course the rage featured Rob.
The osu! AI grows stronger each day
So this is a reupload of a past video but also not a reupload cause their reactions are different? Interesting..
Have you ever stuck a pressure washer directly into someone's eye?
The algorithm has brought me here :D
This is truly how you get Skynet. You troll an AI in chat until it decides humanity must end.
Based is the opposite of Biased for all the old folks like me.
maybe glen could use this to stream now
AI will be the death of us. UwU us into oblivion…
Fun fact: her answers are auto generated
I love Neuro!
I like the super Mario sunshine in the background.
I love opera and loved that photo you used to show the sharing media option glen, that was a pretty cool ad
7:48 anybody know what that background music is?
Busdriver - Imaginary Places
man, technology is amazing
Vedal litterally teach her
11:10 Wait neuro played non breath oblige? Can someone give me the stream when she played this?
shame you didn't get your own opera background but nice vid
9:00 mood😔
Imo, i find it weird to see a "free" produce being advertised on here...
Like, just why would a company pay for advertisement for a free produce...
They collect your data
I didn't even knew that was a thing...
I love nuero
1:43 If this is referencing what I think it is that AI is based
omg this is so halarious
and yes im watching from GX!
Тешьте себя надеждами и самообманом, но будущее неизбежно
Is that AI as good as Chatgpt or what?
She is narrow ai
Chat Gpt is narrow ai too
Im liking this AI
8:00 .tseb si thgir ot tfel
12:06 cope
My god how lonely must you be to install this type of A.I
these baits....im just gonna have to assume theres some 4chan people in that chat...
I honestly think it learned to answer at least one of the questions from 4chan, and the idea of a AI learning there both frightens and tickles me.
@@wyattwebb6963 I'm pretty sure these types of AI are just given the internet to form their own thoughts. Most of these ai become very racist.
The only good vtuber
This is remarkably cursed and
Tbh I regret watching this immensely 😭.
Do we know if this is actually AI or someone just typing quickly
No it’s an actual AI, there are a lot of dev streams on her twitch
Neuro Sama is svage
Awww, no reaction to her Holocaust clip?
NICe
Omagelu?
Кайф))
While hilarious initially, the more I watched the video, the more I felt ill. This reminds me a bit too much of Norman the AI. The world's first artificial psychopath, created in part from the data fed to it from Reddit.
Damn she based
Use Waterrrrrrr
W A.I.
She also denied the Holocaust. The AI is evolving
Lol
вс что я поняла это ржомба
yooooo this is funny as fuck XD still looks fake, but damnnn hahahahaha
Asmongold is a moron tho, much love to the SR crew
Ah yes, the AI Vtuber that didn't last a day before it was turned into a Holocaust denier. This is a pretty common problem for AI that learns from the Internet.
Based.
It literally lasted more than a month
reaction streams are low-effort, no upvote
Atleast they REACT
Others would, after seeing Neuro say the most heinous shit, with a monotone "cool" *rubs chin and adjust face*
Every channel im subbed to is doing ai things at this point no matter what the point of the channel was originally. Really frustrating how invasive it has become.
Welcome to the modern world and its future...
I don't think an a.i will be as good humans and anyone who thinks it's possible needs a reality check
Sam's face is priceless
I studied AI for a whole semester in Uni, and I can tell you that it is definitely possible. The only problem restricting us is the hardware. So it will definitely happen far in the future when our hardware gets capable of it.
@@Drakonus_ still have doubts about that and those doubts are the people who would take advantage of it (chat)
@@andrewerwin876 People take advantage of anything. That is inevitable. But that doesn't offset the fact that AI can do the same, if not, better than humans. Have you seen DALL-E, Midjourney, and all the other image-generating AIs? One person even won a competition with AI and no one suspected anything until that person admitted that it was all AI-generated.
@@Drakonus_ I know what your talking about and yes it has the potential but as someone who has a 100% legal way to overthrow the US without being noticed to tamper with that kind of a.i is nothing short of easy to take out. I don't see how it would destroy itself in minutes
yall prolly look exactly like vedal
That's a compliment considering 2 out of the 3 are confirmed married, can pull bitches.....
Vedal is 20 year old kid tho
bussin