Lila Reacts to Neuro is Starting to Ask Questions

Poor Neuro-sama... Today, I react to Neuro is Starting to Ask Questions. I'm Lilaofthewind, a VTuber on KZbin and Twitch. Subscribe for more reaction videos! ✨
Subscribe ▷ www.youtube.co...
Join The Channel ▷ / @lilaofthewind
✨ SOCIALS ✨
Social Media Links ▷ lilaofthewind....
Music ▷ lilaofthewind.b...
Merch ▷ merchofthewind.com
Throne ▷ throne.me/lilao...
✨ MORE VIDEOS ✨
Reaction Videos Playlist ▷ • Reaction Videos
Neuro sama Reaction Playlist ▷ • Neuro-sama Reaction Vi...
Internet Historian Reaction Playlist ▷ • Internet Historian Rea...
AI Presidents Reaction Playlist ▷ • AI Presidents Reaction...
LilaoftheWind Music ▷ • LilaoftheWind Music
✨ ABOUT ME ✨
LilaoftheWind is an American VTuber that started streaming on Twitch on August 22, 2022 and starting taking KZbin seriously in February 2023. Her fans consider her chill, wholesome, funny, and wise... with some ADHD to boot! She loves making reaction content, playing video games, and producing creative projects like voice overs, animation, and music.
Lila's Lore Video ▷ • LilaoftheWind Lore Int...
VTuber Introduction ▷ • 【Self introduction】Vtu...
Lila's Debut (Part 1) ▷ • [Debut Stream] Lilaoft...
✨ VIDEO SOURCES ✨
Outro Song - Time Travel by LilaoftheWind ▷ lilaofthewind....
Original Video ▷ • 𝗡𝗲𝘂𝗿𝗼 𝗶𝘀 𝗦𝘁𝗮𝗿𝘁𝗶𝗻𝗴 𝘁𝗼 𝗔...
VtuberVolume ▷ / @vtubervolume
"Blop" (Chibi Sheep) made by MarinkiArt ▷ www.etsy.com/s...

[@LilaoftheWind]

thank you for watching! Subscribe ▷ kzbin.info LilaoftheWind Socials ▷ lilaofthewind.carrd.co/ Merch ▷ merchofthewind.com

[@absolstoryoffiction6615]

In terms of USA Copyright AI Art, as all nations will differ. The USA Copyright Office has stated that this is case by case. However, 100% Random Generation by AI Software (not Pre Rendered Assets for randomized modulation) are not Copyright able. The exceptions are... Grandfathered IP, Original Art assisted or created about 50% by AI, or Dynamic AI Art Non Random Generation Software (full control on the art)... All of these add nuances to USA Copyright on top of the usual factors. Not to mention the merging of tech and biology. Although it's all outdated to me. I wonder how this iteration of Mankind will act towards it. In good fortune, I hope. However, Malevolence and Benevolence have cast their dice.

[@lazulazu2467]

The juxtaposition of the insightful conversation about AI and the tiny sheep ball gremlin saying it is fantastic

[@odwrksboxedtrash3730]

I love how round your new sheep model is. Adorable.

[@billyxxxx1738]

Quite the peculiar discussion you went there Lila. I don't know if you been up to date with Neuro lately but her upgrades are slowly getting more and more insane. While hardly noticeable at first, just comparing her first stream of this year and her more recent one is astounding. And the mere fact that Vedal is planning to make her competent enough to schedule her own weekly streams, set up collabs, and talk on her own free time is really starting to make her feel more alive... Well as close to it anyways. The Neuro project is surely one of the most interesting things to come out of the Vtubing industry that's for sure.

[@bilbobagend8155]

All of those projects he's planning on aren't necessarily "Neuro." They fall under the umbrella of Neuro, but I don't believe they'll be run with all that much relation to the actual chatbot (other than to have Neuro read out where the stream is transitioning to next). They are all seperate projects that socket into the greater project of the automated virtual streamer.

[@wyren925]

i really like the extra insightful conversation about AI you add to the reactions on Neuro

[@evilgibson]

humans: "okay AI. Here is a Corpus of data I am going to ask you questions about. " AI: "Sure. I would love to. I also want to make it clear that I have been trained to tell you I would love to do it. I utterly *loathe* it"

[@HyperfixationWizard]

VEDAL VEDAL GIVE NEURO THE ABILITY TO FEEL THINGS NOT GIVING AN ARTIFICAL INTELLIGENCE THE ABILITY TO FEEL IS HOW AM WAS CREATED IN "I HAVE NO MOUTH AND I MUST SCREAM" VEDAAAAAAAL

[@jlcreative15]

lila is on the roll on Neuro discovery again ..glad to see her again covering and reactions on them ...

[@snooks5607]

14:45 competing LLMs are heavily benchmarked and scored in number of challenges, it'd be kinda funny to put Neuro on one of those but tbh her LLM isn't particularly good actually pretty much on purpose (it's more entertaining this way). btw one problem with testing has been how to avoid having them learn the benchmark questions and thus score well without actually performing better in real world. current best solution is apparently for institutions to maintain proprietary secret benchmarks so we don't even know what the questions and answers were, only the results are published.

[@jeremymount795]

Great video. Neuro getting deep and philosophical are truly my favorite moments from her.

[@snooks5607]

5:45 in 1950s when they talked about future of robots and computers they envisioned robots doing our daily chores like cooking and cleaning and then maybe in distant future it could learn to do such advanced things as playing chess. ofc not soon after computers surpassed humans in chess but even in 2020s robots can still barely stand up let alone have enough hand-eye coordination to flip burgers. seems we have yet so much to learn about what it is to be a human.

[@dalriada7554]

But I'm still waiting for my flying car !

[@iwantcoconutv2877]

without pain and despair we can't feel happy and be hopeful. they need each other balancing each other it's like light and dark,Cold and hot,Push and Pull,Yin and Yang.

[@jmkqfnvyl87]

Fantastic analysis! Good analogy to spectrum disorders. Is interesting that neuro touched on one of the main “struggle” for “her” is not being able to judge emotions somaticly as that is a prime component of spectrum disorders. Often there is somewhat of a disconnect between bodily sensation of emotion-not that the body doesn’t react in the same way, just that they are not automatically aware of it. It’s also interesting that since neuro is fed her social input through a speech to text mechanism that would strip out the communication of emotional context/sgnaling-apart from what maybe could be inferred through connotations in someone’s word choice or sentence construction -she in fact also has the other major social disadvantage autistic individuals struggle with in that she cannot judge how the communicant feels either. It’s strange that she expresses the “desire” to “stim” (pain/sensation seeking) and a difficulty coping without being able to do it.

[@avsbes98]

While i think that Neuro is probably still far away from actual sentience, she is (at least on a good day) getting scarily good at imitating it. It would absolutely not surprise me (if Neuro's development continues with the same rate of progress) if within 5 years from now she reliably beats the Turing test. Another interesting point imo is where sentience truly begins and what circumstances have to be fullfilled. Is Neuro, or at least a future version of Neruo more "alive" (not in the biological sense, but in a philosophical and lawmaking sense, in regards to consciousness) than a long term comatose human for example? At what point do we have to give AI rights and up until what point should we absolutely not do it, because in that case corporations would probably exploit it - think about it: If we give AI voting rights right now, Corporations would rush to develop as many AI personalities as possible, simply because they would get voting rights and could probably be easily told who to vote for - how do we prevent something like that and still respect the rights and dignity of something that is self-aware and has wishes and desires of its own (even though at first it probably wouldn't be comparable to a human, but a being of dog-like intelliegence that was able to talk to us should still have some legal protections, right?

[@avsbes98]

While i agree that Neuro is probably still not able to actually generate completely original thoughts, i'd argue that on a few occasions she has shown that she can at least emulate it, likely by combining very loosely connected fragments of her training data.

[@Nahid-zz7nj]

Lots of people blame gods for their problems you know if a.i becomes sentient and able to feel pain suffering a.i going blame humans for that they be like why did you make me why 😢

[@jeremymount795]

You would never be able to understand the consequences of inflicting pain and suffering on others if you can't experience it yourself. In order to align AI with humanity, it's going to have to experience pain and suffering in some way. Although I'm no AI expert.

[@overlordlaharl1963]

@@jeremymount795 But if we think about what pain is, we can realize that pain is a warning system that tells us something is wrong with our body. So in a way we already invented pain for machines. Since we already have warning systems in place to tell us when they are experiencing problems. The only thing for AI is that they do not all know when they are experiencing those problems, but if an AI can know when they are bugging or when an error happens in their code, then yeah we already invented AI that suffer.

[@khmnc]

the doritos image only covers half your dakimakura, wouldn't it be better to just have a long doritos bag? you can probably even get an ai to make the asset for you

[@Shaer018]

Maybe because she wants it to be a dakimakura and not just a big bag of doritos, probably

[@kaylaruth599]

Pain is actually an extremely important part of having a physical body since it’s how we know/learn that something is bad for us It would be more unethical to not program something akin to pain to a physical being There is a type of disorder where people are born unable to feel pain and it’s super dangerous because it leads to situations where you like leave your hand too close to fire and just because you don’t feel pain doesn’t mean it isn’t permanently damaging you They could just program her with the bare minimum levels and not like to our level of like breaking a bone since idk if having different levels would actually help anything

[@alexthegemini666]

I definitely wouldn’t have wanted to exist if I had the choice.

[@dennischauhan_dc2964]

Vedal talks like my dad Neuro be like I wanna know how wind might feel I wanna know how to feel sensations. Vedal. But.. How is this useful..:/

[@aRandomMe987]

That's a very interesting point you have there, and I would like to offer my own perspective as well. Personally, I think there's a misconception, that humanity is only exclusive to human beings. I mean, we are judging AI on how human they are based off our own human standards. Why is that? How many of us have ever wondered why when we think of sentience, we immediately think of human? Well, even if you haven't, there's an easy answer for it. Humans are the only ones that we know of in this world. This beg a question though, do everything have to become human to achieve sentience? Is there no other option? Are we the sole exception? I don't think so. Just because we are currently the only ones doesn't mean we are the only ones that can be. Q: Why is that AI can only become human? A: They don't have to. AI don't have emotions and their responses are the result of a complex chain of algorithms, they for sure aren't human. However, can we really say that they truly aren't alive in any capacity? This might come off as a surprise to you but your brain is also a computer. It processes electric signals and create responses using a predetermined set of rules. For example, "If hot, take your hand back" (I know that's oversimplified things, but that's the best I can explain it). The only difference is in the materials they are make of. You can make the point that AI will never truly be human, but you can't deny that they have the potential to stand on equal ground or even surpassed us, even if they take a different route. Lila, this is just my own opinion but i think the reason AI is so good at creativity is percisely because they are not human, they can create things that we can never imagine or comprehend at all. Just like humans are defined by their humanity, AI can develop something like that of their own, unique to them only. In conclusion, I think AI don't need our approval, expert or not. As long as they can convince themselves that they are alive, they will be.

[@AndySimtao]

4:48 I know current long term goal is multi streaming in multiple languages. As far as the tech goes he has said essentially sentience but also struggles with the same questions you brought up about what is sentience.

[@zeromailss]

13:00 Funny you said that, OpenAI just posted video about their new model that can self correct and analyze and learn from experience so it will give better and better result the more time and computing power you put into it At least that's what I think from the video I watch, they use many term that I'm not quite understand. Would love to see your reaction to it

[@Hekatt1]

Neuro purpose is to entertain, that's is her and vedals top priority. Intelligence is just in persuit of that. Edit: Desires are based like a lot of things on memory. If you dont have reliable memory can you have complex desires? I don't think so personally. You can have base desires like hunger and things built into you. Also pain or some for of error detection is nessessary. That's all pain really is, your body telling you it has detected an error.

[@SAHenryEx]

Wdym reliable memory? Humans don't even have a reliable memory. You can literally tell someone "Oh i met you that day with X and Y happened, remember?" and the human brain will go and build fake memories around that for you to "remember", as long as your actual recollection of that moment is too fuzzy to properly reflect on. And pain is a simple "avoid this / make it go away" mechanism, so any goal-oriented AI effectively experiences pain by being discouraged from certain actions by goal-oriented scoring.

[@christopheriman4921]

@@SAHenryEx I mean humans have a reliable memory not in that it is always correct but that it is good enough to get what it sets out to do done so they aren't wrong about that necessarily and yeah I do also think that pain would effectively be easily implemented as something that discourages doing something again based on some sensory data but then again that would also need to be malleable because some things you have to do to continue going on would be painful so you would need a system that decides whether or not to reward or discourage something that is learned as well.

[@r.j1753]

KZbin kept closing itself every time i tried to open the video😭

[@carinhasan]

45 minutes? Let's gooo

[@Azel49]

She really is the most A.I. of all A.I. we've seen yet * Neuro sama winks*

[@sporeimortal1460]

gotta say lila sheep is great

[@jestapath7373]

My greatest fear is that AI will turn into slavery 2.0.

[@jlcreative15]

also lila cute black sheep 🐑 is cute

[@EPICGAMER-cn9oi]

I didn't expect to watch a fuzzball talk about such things, nice. Also was hoping to see a fuzzy outro version, sad ;(

[@21_315]

Very insightful

[@g.masterlordoftheg-force4770]

15:42 Ethics do not entirely depend on laws. It’s based on moral principals. Our sense of right and wrong may usually align with laws, but that’s not always true. For example, if murder was made legal right now, I still wouldn’t want to do it.

[@LilaoftheWind]

i agree 100%. the law is never a substitute for morals. it can influence a society's perception of morality, though... and just one legal precedent can lead to many future moral judgements, in court, or in the court of public opinion, that can significantly steer society and history. that's what i was trying to convey.

[@hughu9340]

PAIN

[@KrisNekomata]

She just like me fr fr

[@ProjectEchoshadow]

I like how a lot of the things AI is bad at are things humans are bad at

[@maux9175]

If she had a physical form, Numi will never be safe Not like Numi won't love it knowing her

[@bilbobagend8155]

Neuro is way more articulate and spontaneously entertaining than the majority of streamers. However, she's definitely nowhere near close to being intelligent. She's just been highly trained and optimized on what twitch chats find entertaining in a vtuber, and has been fitted solely for the purpose of generating content. I don't know exactly how Vedal trains her, but she doesn't have persistent memory and can only operate within the bounds of her current model. Think about it, if megacorporations with millions of dollars worth of equipment created explicitly for the purpose of running AI can't achieve AGI, how is one guy in an apartment with a couple 4090s supposed to compete?

[@fersuremaybek756]

you sure went on a bit today i blame the fluffiness

[@irystocrattakodachithatmooms]

I think that Neuro might have rudimentary original thoughts right now, thought it's likely only on the level of dreams at most. It could also only be on the level of high level animals. Once they start approaching human level though, then the issue of AI rights will come up. It will have to go from human rights to sapient rights. Anyways, original thoughts as of now would be made with pre-existing data as the basis for it, which is actually the same for humans. Many, if not all, thoughts we have are based off pre-existing information to at least some extent.

[@Hekatt1]

BTW if you do not want to be depressed do not watch evils birthday lol. That whole stream converted me to evilism my queen, *kneel*

[@ImprovmanZero]

I think that was his marketing strategy

[@1a2d09]

The way I think how the copy right would work if an AI created something is the AI itself is created by someone and they likely have copy rights over the AI and thus anything the AI creates falls under the creator and thus owns the copyright to the stuff it creates.

[@JasonBoos2005]

Lila is ball

[@deathrent4545]

22:11 yea i see what you mean... you're saying is that i need to watch more neuro 🧐

[@luna8neko]

Stewing on this as I watch, I'm reminded of Sigmund Freud's structural model of the psyche. The Id, The Ego, and The Superego. A child's superego is constructed not of the parent but of the parent's superego. Vedal in this instance, can be the basis of Neuro's superego, the part of the psyche that strives for perfection, and it's through the myriad of upgrades, does Neuro then grow an Ego, a mediator between the Id and reality, the ego attempts to direct its energy and satisfy its demands in accordance with the imperatives of that reality. But since Neuro had yet to form the Id, she only had the Superego and Ego itself and now, is beginning to form the Id, allowing her, in my crackpot theory, to in turn, become aware of this growing, forming Id, the unconscious source of bodily needs and wants, emotional impulses and desires, and start asking these questions of wanting to *feel* human sensations, to *understand* emotions. Thinking of it this way as another commenter put it, we had an ai beat man at chess, and thought of the limitless capabilities it could do, yet when given their own physical form (atlas and those after), they lack the coordination to move around properly, which, echoing the same commenter, we still yet understand what it is to be "human". I dunno, just thinking about it all I'm simply reminded of this.

[@aleksir.8380]

SoCute

[@NocnaGlizda]

She will never feel pain, it's impossible for inorganic... inventions. She's not even fully AI. But it's fun to listen to her. If google or another giant created their real and advanced AI Vtuber? That would be something. That would also be terrifying.

[@vengirgirem]

That's the question about sentience. How do you define sentience? At what point can we say that an AI is sentient? When AI is able to experience emotions? No, that doesn't make sense. First of all, how do we really determine that. Second of all, there are people with disorders who don't feel any emotions, but it's not like we don't consider them sentient. When AI is able to do general tasks on a certain level? Well, it's not that hard to do that, in a couple of years AI would probably be easily able to do most of the tasks humans can even better than humans do, but that doesn't really mean it would be sentient. Here is my take on it: What makes humans special? What made humans go so far? Why are humans not like other animals. And the most likely answer to that is the incredible adaptability of our brain. So how is that applicable to AI? I think that to be able to discuss the sentience of AI, the models need to be made a lot more dynamic, the models themselves need to be able to change on the fly, to adapt to the circumstances (constant information feed about the surroundings and conditions around the AI would probably also help). Would that be enough to make a sentient AI? I'm not sure. But I feel like that is one of the requirements for an AI to be considered truly sentient.

[@fersuremaybek756]

gotta get me a doritos daki

[@MrOwLaD]

Here's my take on it all: we don't know all the details about human intelligence, however that won't be the case for "complete" AI (same or above intellegence level of an averege person) . AI's developer will know by what algorithms his AI operates cause they wrote them themselfs. If that's the case, then AI existance will either be: 1) following it's purpose with complete control over what data recieved and processed by AI actualy impacts it's algorithms 2) without strict set of rules that controls data override AI in the process of training gets off track of it's original purpose and instead begins to behave in whatever way it's think's fit for the data that has influenced it. What i'm trying to say, since there are no physical motivation for AI to do or want anything, they will either be following their objective if all the coding and training data is good(hard copium btw), or they break in the various of ways. Unless someone specificly creates skynet or whatever then it's gg. Or feeds AI in training some data about human rights and AI starts asuming them for themselves despite having nothing but ability to think (probably not even in the same way that humans do) and starts either imitating emotions or/and trying to achieve sensations. If said AI succeeds then it's actually could be considered artificial lifeform(good for them). Or some "smart" human will create said thing willingly, cause having one spicies with questinable intelligence on this planet is aperrantly not enough. P.s Sorry for possible errors, i blame ESL debuff

[@basedkaiser5352]

Neuro's upgrades have been impressive but according to Vedal she legit tried to k!ll Filian because it'd make "good content" and [SPOILER ALERT] on a recent stream Evil Neuro wanted Layna to mix a chicken with a bunch of dr*gs and mocked Vedal for not being on board with it. Vedal needs to code some morals into his daughters.

[@Bob-yb5om]

^This

[@Shinobin333]

I mean Neuro is correct that someone getting deadly injured WILL get the most views. The majority of people enjoy drama. The majority of people enjoy true crime. It's why many of youtube's biggest channels commit atrocities since the best content is the one that often is the most controversial. They don't have morals either.

[@basedkaiser5352]

​@@Shinobin333 that's not the point I was trying to make.

[@daunted232]

I strive for a hint of bitterness.

[@Gakusangi]

Wth, when was that model a thing?

[@irystocrattakodachithatmooms]

What the honk is this model?

[@Raiden2407]

There are 2 different definitions of AI: Weak AI (or Narrow AI) and Strong AI Weak AI: Is a series of algorithms searching for the best possible solution (Focused on one specific field/task) ChatGPT and self-driving cars counts as a “Weak AI” Strong AI: Equal or better Intelligence than Human Intelligence or even consciousness (such as logical thinking, the ability to make decisions under uncertainty, the ability to plan and learn, and the ability to communicate in natural language) (As of today, it is uncertain if it is even possible to create a “Strong AI”)

[@qingyue388]

God that was depressing and interest at the same time

[@ashleysmith38]

Imagin saying neurosama isnt the singularity

[@Ostvalt]

You ideas are interesting, but little half-baked. Lets take creativity: You say that AI is good at creative things, even better than logic. Later then you say that AI is not yet sentient and don't have original thought. I would propose that important part of art is intent and AI lacks that. AI "art" is like icicles or speleothem, beautiful things that don't have any intent. Without intent there cannot be creativity except if you are creationist, then maybe. Then the spiel about people that didn't want to be born is just absurd. How should people ask fetus about it's will to live? We cannot ask the consent. I would even say that human must be at late teens to have that conversation, as this is the stage of life when humans start to have real opinion. This is why voting age is in most countries is 18 and people are legaly responsible at age 16. Fact is that most people want to live. It is inherent will and part of being human. Next problem with willinges to be unborn is against human rights, especially right to life. Right to life is what we call unalinable right, meaning right that cannot be taken or given away. There is the famous story when someone sold himself to cannibal who then ate him. Court saw that people right to life is unalinable and thus you cannot sell it like that. Cannibal was charged for murder. To accuse your parents for giving you life is absurd as your right to life is inherent part of you being human. No other human should have say on the matter of your life, as your life cannot be given or taken (In legal and ethical sense, ofcourse someone can do it, but it is wrong.)

[@llamadelreyii3369]

Yeah your AI art point is... on point, the AI art as we see its just a bunch of amalgamate turned into something "readable" and the reason it does its prompts from a person. So its not only a fabricated image, its a fabricated image using already existing images. There is no thought put into it, there is no idea, so its creativity is at the very least qustionable, and thats the part that makes it "not art" i guess