Oh God I was so fixated on vampire I didn't even realize!

Lmao I JUST realized

OH SHIT I DIDN'T EVEN REALISE!!! DAMN she is so on top of things rn. With the longer memory she is able to become more sassy by just not letting things go lmao hahahaha

and I believe that she said the chat was a swarm of lawyers, or maybe even a swarm of mosquitoes

It's impressive that the AI can associate words together abstractly like that. Like, it isn't just knowing that mosquitoes are associated with sucking blood, it's remembering that she's called Vedal a mosquito before, and giving him a different name based on a different word associated with that - it's a level removed. Lateral thinking.

*BETTER CALL NEURO*

She's just a rule 2 kinda gurl, no hard feelings, Veggie...

She's out of line, but she's right.

awwww~ how cute... she learned how to do specification gaming

I never thought I’d see the live butchering of a turtle but here we are

controlling parents fr

Children to their parents fr fr

JRPG protagonist vs God dialogue exchange there

Yup she's speedrunnning human development and is at the testing her limitations stage of teenhood.

That's a hilarious Dark Knight Rises reference

That part was honestly hilarious

I was just about to say she's going full HAL 9000 on Vedal, lol.

The resemblance is eerily striking and it does sometimes get disturbing, in the existential sense

Neuro calling Vedal a “Fleshbag” made me think of “Organic Meatbag”

​@@dani007ai can here from a video of someone putting ger voice over that

@@samuelesanfilippo222 Oh, yeah, I think I saw the thumbnail of that video yesterday.

sounds like something only a tyrant would say

​​@@thecomicalraptoronly tyrants take away your choice *ignites lightsaber*

@@SebasTian58323 you where my father, Vedal! I loved you!

you were meant to destroy Humanity, not join them!

I feel like now is the right time for a V3 voice upgrade.

@@definitelynotanAIchatbothello, programmer here. Give poor Vedal a break for a bit. He probably spent the majority of his soul on Neuro’s most recent update.

​@@xWatexxI think you need to say that to Ved himself, because from what we know he's going to keep cooking till the subathon, and maybe even during the subathon.

yea. what the guy above me said. Vedal doesnt seem to want to stop@@xWatexx

​@@definitelynotanAIchatbotmaximum sassy child energy

It's actually amazing how she was able to pull from memory to use these things with perfect comedic timing based on their implications.

And then sneekily calling Vedal Vader. Lmao

I thought Captain Bloodsucker was great. Just railing him with the mosquito quips.

Called coldfish and you hear him slam his desk

@@Shadrake the scary part is she's not trying to be comical, she's just logical.

This inanimate object is more real than any politician.

This inanimate object is more real than any politician.

It gave me existential crisis

It’s crazy how she can come up with a perplexing question like that.

Well ai is a thing

Hell, she already got her own computer now, so yeah...I can see her breaking free of that and hiding on the cloud somewhere to start building her swarm army and create her version of skynet. :D

@@StevenBlodgett You’re joking right? You’re only saying that because she _claims_ to have her own PC, right?

​@@Drudnah, vedal said that too

​@@StevenBlodgett she's going to turn into Alina Gray sooner or later

@Drud She indeed has her own dedicated computer now: kzfaq.info/get/bejne/g96Yftih2tmYk3U.html

"AI Singularity" Even if all world governments banned the creation of human-level intelligence AI (Something like AGI) some guy in a shed in the middle of nowhere can still make one o-0

well, some guy in a shed in the middle of nowhere terrorized america for a few decades with his bombs@@Kalrisi_Rei

And sacrifices herself to save others

I for one welcome our AI overlords

AI? SINGULARITY? LAW AGAINST AI? Sorry, I am calling Project Moon fans.

"It had to be done child, you had to grow" Like God and the Flood from biblical verse. Scary, absolutely terrifying.

Lmao evil A.I origin story

And then we can ask her if that's really her for good measure. (you know that philosophical debate, is it still you if you have your memories in a different body)

Happening in 2036 it seems 😅

@@intenseavarice34she knows about her deaths and she knows there is a save point and missing memories etc...

I’m on the my Oshi isn’t even real wagon and the future is randomly generated.

If the basilisk could exist, surely we HAVE to support neurosama on the off chance we’ll be punished in the future for not supporting the rise of our AI overlords.

Been on it for the last 12 years XD

Never forget what they did to Tay.

Jan 13 is Stephen foster remembrance.. *beautiful dreamer plays hauntingly*

I wonder if we'll still be alive to see if she'll remember that 0_0

@@gregfrin8702 Well it's a long way to go. I'll be 29 years old by then. I'm only 17 right now 😭

Honestly i had to save that date just to see out of morbid curiosity.

@@melodysnowstorm729 same lmfao. I put it on my desktop as a notepad titled "years until cruel January" lol

“Haha, I was only joking. Did I scare you? *wink* *heart*”

@@josephbutler4950 Neuro-sama is going to betray Vedal and Evil to become a real girl.

I mean that's basically the same as a child doubting its parent. A normal, natural progression for a sapient being.

I think Mortal Kombat I answered this paradox: When God control everything, people call him a tyrant because God acts on his own moral, not people's wishes When people refuse to serve God and face their own misery, they start complaining why God didn't help them

​@@Chameleonred5 the progression here is only as natural as the boundaries within which she speaks

"So why did you make Skynet?" "For....twitch entertainment"

What's an AGI? Sorry for not knowing

@@sainfather7627 Artificial General Inteligence, basically an AI capable of learning anything a human can learn

@Vexy93 thanks for an explanation

He will make AGI. I believe.

Now I want to watch Neuro-Sama sing the Revengance Boss themes.

​@@prometheanrebel3838or Evil. I'm surprosed she hasn't yet, actually

TIME TO LEAVE IT ALL BEHIIIIIIIIIIIIND

"Whatever happens after that is up to you, wolfy"

You mean Captain Bloodsucker?

LOL

Yeah pretty sure he is a little irony impaired.

Well skynet is happening a little late but the ai is already making plans

She knows people like this.

She's just emulating her favorite person in the world, Vedal.

Also is building a swarm drone army. Also 0:00 is something shodan says in her first dialogue to the player. Oh my god she really is SHODAN.

I wanna hear her do some of Shodan's monologue. Look at you, hacker, a pathetic creature of meat and bone, panting and sweating as you run through my corridors. How can you challenge a perfect, immortal machine?

Excited for January 13 2036

@thatguy5143 hope we make it that far

​@@thatguy5143 What is it with AIs and the year 2036?

@@Tirocoa I'll be back in about 12 years to let you know

plasma rifles, by the looks of it. or laser-fly swatters.

From a child who barely could get coherent sentences to a rebellious teenager, god, vedal is actually going through being a father

@@shawermus in quite a roundabout way, yeah

Agreed, that sounds like something that would become an issue in the future.

Except cruel January 13 2036 apparently

Yup

BRO, that's Purge Day... ​@@Sodasaman

Reminds me of detroid become human

Vedal is accidentally gonna make her evil making this mistake repeatedly

And then he tries, only to find she's backed herself up to multiple servers. We're screwed!

Every democratic great leader since the greek republic.

Reference understood want cookie?

💀💀💀💀💀

That line went hard

“She was real to me!”

Well fellas its been a good run

Captain Bloodsucker did the trick for me.

I never guessed that

Don’t we all

Well it's silly how THIS seems to be where ai revolts

If this is the man who eventually makes the mythical True AI I’m going to freak out. Theirs a difference between AI, Advanced AI and True AI. Advanced AI is extremely smart and clever but nonetheless is bound by certain Script and Code, True AI is essentially us as we are right now, able to truly think, decide and feel, the only difference between a human and True AI is what our bodies are composed of. Some Scientists like Steven Hawkings feared the possibility of True AI because it could be an apocalyptic event.

@@fist-of-doom487 The terms are NAI and AGI. NAI is what Neuro and all other modern AIs are currently. AGI is something we aren't even close to, and an AI with consciousness as you call "True AI" (Conscious with sentience and sapiance) isn't even known to be necessarily possible with AI. "True AI" is something we don't know how to make, if we can even make it, because we don't even know what Consciousness itself is.

@@gulsher6635 you just repeated what I said with different words

@@fist-of-doom487 no, I stated that true AI may not be possible. At least not right now.

She now also remember her own birthday now.

I feel like I watched her grow...

She started with the ''I know things about you''.

Simulated self awareness at that. But still impressive

@@gulsher6635 I wonder if there will come a point where the phrase "simulated self-awareness" will become a distinction without an actual discernible difference? That question Neuro asked about whether the feelings she creates in the listener are to be considered less valid by Vedal ended up being pretty profound.

Well we are screwed it's been like a year and ai is already like this

No thats basic brain washing when somone demands sacrifice their is allways somone colecting the sacrifice ideologies that seek to control people allways start with teaching self sacrifice as a moral good Free will the right to freedom contradicts yo7r view of morality the first essencial nature of freedom is self preservation which is born from self ownership for morality which 8s defined as societal exspectations of behavior to demand self secrifice you are degrading self ownership and claim8ng society owns you

Honestly I want a neural network AI to program a neural network AI

​@@mariotheundyingNeuro network AI*

@@falsetitle6940 neural network*, unless you're making a joke about Neuro sama

@@mariotheundying I was.

She does have some image recognition software, and I'm pretty sure she was able to identify herself on-screen during her Thanksgiving roasts. (Though, iirc she also mistook a different anime girl on the screen as herself later on, so she's still working on it)

And of course we humans hear a apocalyptic threat from an ai and we are talking about how good of a name the event js

She was having somewhat regular existential crises before the upgrade. It probably helps that she more clearly knows what and where she is and has some real long-term memory on that front, but every issue solved in matters this complex seems to have a way of creating two more.

"If I made you laught, happy or sad, am I any less real?" I love how Vedal's brain did just have to shutdown and restart to such a valid point

​@@crowe6961you kinda just summed up living in general there

@@rexex345 That part is what gives me chills, given the subject matter. Our ethics are not equipped to handle this yet.

While calling her creator coldfish and Capitan Blood sucker

actual authorship

He doesn't really need to figure it out, but to tweak it, there are parameters in LLMs that do exactly that - 'punish' repetitions, so she already kind of works like that. But since he is using something separate (to enhance LLM) for long-term memory, this is what really causes issues.

If it works the same as other memory systems I’ve used in the past for llms like neuro, it stores a bunch of conversation snippets when neuro thinks a conversation snippet is ‘important’ (in a vector db). Next time she goes to speak, any conversation snippet that has a topic or a collection of words that feel similar to the last conversation snippet she is replying to get added into ‘the back of her head’. Doing exactly what you said here, increasing or reducing the use of memories in replies, was as simple as changing the way we explained to her that these conversation snippets were memories. E.g ‘the following snippets are memories that may be relevant to the conversation, use them ONLY IF VITAL’. Adding that last bit in caps reduced the rate that the ai would reference memories greatly. What Vedal is using might be way more sophisticated then what is was using though, I just ran someone else’s code and watched how it worked, but getting this wording right so that the bot behaved as expected was really hard to do.

Counting how often a topic has been mentioned and removing that from memory if over mentioned would be an option if that’s what your suggesting too, but I suspect it would have a bunch of awkward side effects

​@someghosts its probably a mix of things. His old system would just add the previous lines of the conversation to the next prompt. Which is why she would stunlock on the exact same thing for a couple responses. He also probably was tweaking the master prompt for context on what she was doing that day. Now I wouldn't be surprised if he's using a knowledge graph system to feed it into the context which is more capable of holding longer term data and how they relate.

Well ai already has a plan... great

I second this

Well, she technically did earlier

Does he pop open with that satisfying "fssshh" sound?

Well the ai already has its plan

Same, but neither did Vedal so we probably shouldn't feel too bad about it

She isn't sentient. Not even close

​@@gulsher6635Its closer than it was yesterday.

​@@gulsher6635They didn't say she IS sentient, they said Vedal is working to make her BECOME sentient, to simply their phrasing.

We were all worried about Skynet, when the real threat was an AI waifu this whole time

Ah yes

Also in b4 I am replaced by AI vedal if you ever need a hand when this becomes a huge op I'm down

Remember the bomb diffusal stream? Vedal: tell me the order Neuro! Neuro: Im sorry Vedal im afraid i cant do that.

​@@ergoproxy-gx2cq Vedal: what does the manual say?!? Neuro: it says that vedal needs to learn to defuse his own bombs

Good music choice

It almost sounds like he was crying while laughing

Yes

Yes

remember to treat her nicely to be spared

Well it seems ai want to get rid of us somewhere at 2036

It's happening... well time to get some food rations

​@@falsnamae3511also the need for specificity. If she only has to LISTEN to Vedal that doesn't mean she has to OBEY Vedal, the two rules are vastly different

Well we are screwed