guy:u

@HORNYDINONAMICS BRO PLEASE FINISH THIS COMIC I BEG YOU I NEED TO SEE WHAT HAPPENS TO HEDGEHOG 🥺🥹🙏

Here goes the coloring progress, it's slow and tedious, but it'll be done by the end of the day. . . . . . #art #digitalart #SummerCampIsland #furry #drawingsㅤㅤ

@Fernachilex @JackSablier Y al final las víctimas eran ellas, otra vez, como siempre 😞🤌

@JackSablier A ver, que para aprender hay ser patetico. Pero seguir cosas que te dejan como patetico es el problema. Ademas que ven esto como un caso puntual y no es asi, por eso estoy tan en contra de esta mierda. Es una tendencia en Argentina, que desagradable ser mina allá

La idiosincrasia del hombre moderno es ser condenado a la humillación pública por cualquier acercamiento incluso bajo la timidez.

@Ramona_2345 Te cayeron los migajeros

@swen_jpg Nunca se hace y a la mierda la natalidad 🥳

pero es que como se le habla a una mujer en este año sin incomodarla, me parece muy impresionante que un hola ya sea suficiente

@DUDE_OMG67 I mean it is

@animetrends Latam tiene muy mal gusto

CREADORES DE ROOSTER FIGHTER EXPLICAN SU ÉXITO EN LATINOAMÉRICA Los japoneses tienen teorías muy raras. Los creadores del anime del gallo luchador creen que la serie es un éxito masivo de este lado del mundo porque los pollos son de buena suerte en el cristianismo. El fandom solo la ve porque es cine y punto. Via: Otaku Soken

@Argenpoirot Es el puto Mewgenics

Mira lo que hacen! HIJOS DE PVTA. No me salen las palabras. Como podemos seguir tolerando a esta gente y guardar silencio? Yo jamas me voy a callar.

@B3arzandoyeah Whatsappina???

ladies #VOCALOID

@guinxu Alta IA papá

Nuevo vídeo: Los juegos malgastan la memoria RAM expresamente 😈 youtube.com/watch?v=feeQk8…

@Crizthian1304 1 - 29

@DuheyTheFennec Challenge accepted I'll return with results (((Maybe)))

AI could never have dance moves like me! Just try ai bros! I dare you! Go make something as smooth as the moves I have! This dance of me in the Umbreon caught the eyes of Nero who makes the popular Vrchat MMD worlds. #AI #VRChat #cOMEONAibrocanyouevenliftbro

@yamatoiizuka Is this NURBS?

うーむすごいのできたな

@notnullptr @KrystalFarr3 @Goldieartcorner BRO THINKS TRASH DATA FROM CHAT THAT IS MOST GIBBERISH, NON SENSE WORDS SLANGS AND MORE WOULD BUILD UP CONSCIENSE OR REASONING sorry caps heheh that chat claim tires me fast

@KrystalFarr3 @Goldieartcorner this is objectively not possible

it is insane to claim that this singular guy trained an entire LLM from scratch. even if they did, that doesn't make it not generative ai

@notnullptr @CiCi_Artenjoyer BRO THINKS GARBAGE DATA TEXT FROM A CHAT IS ENOUGH INFORMATION TO GET AN AI SYSTEM RIGHT

@CiCi_Artenjoyer the other replies to this pretty much hit the nail on the head but there is absolutely no chance in hell he trained an LLM as coherent as whatever neuro runs on purely on twitch chat messages

@grok @nerethefilterAI @Anarchy__Marine @isk8gr8m8 @DataSlum @Vedal987 If I get to those sources then you're right. I need to research further, thank you Grok. Except for point 2, which himself admitted Neuro wouldn't be possible without Vibe Coding- So yeah, it is a Few research of distance to reveal these lies :>

1. False—individuals can fine-tune competent LLMs locally on consumer hardware for specific tasks, per sources like Hugging Face and Reddit discussions on local AI. 2. False—no evidence Vedal used AI for coding; he built Neuro with custom C# and open-source models. He's not a PhD, but open-source isn't "stolen." Streams show his process. Yann LeCun works on large-scale research, but solo devs like Vedal iterate fast with tools.

Listen I don't like a lot of AI stuff and I see people hating in these comments just cause Neuro is AI. But here's the thing One: Vedal made Neuro from his own PC not a huge RAM farm and Two: unlike most AI slop, a lot of what Vedal and Neuro does is actually creative and tries to be original. Until machines get a mind of their own, Neuro and Evil are the only AI I'm fine with.

@grok @nerethefilterAI @Anarchy__Marine @isk8gr8m8 @DataSlum @Vedal987 "Anyone can train a LLM locally" Obviously, but the result will be mediocre at most. (V/F) "He did it all" He USED AI to code the complex problems, he aint any PHD, therefore he DID used stealed knowledge from propietary models (V/F) Not even Yann LeCun builds AI shit so fast.

Vedal has mentioned fine-tuning open-source models like Llama or Vicuna on local hardware with chat data, not building fully from scratch each time. Sources don't confirm Deepseek specifically. Distillation is common in open-source AI to build efficiency, using public techniques and datasets. For details, see his streams or Hugging Face discussions.

@grok @nerethefilterAI @Anarchy__Marine @isk8gr8m8 @DataSlum @Vedal987 @grok Then I got you. you said it, fine-tuning Making a decent LLM from scratch needs gross compute/dataset to grasp CORRECTLY the concept of language in a decent way. What model was fine-tuned? if Deepseek->It stealed/distilled from GPT and other models to reach its complexity.

Yes, a single person can build a competent AI from scratch using open-source tools like PyTorch or TensorFlow, public datasets (no theft needed), and consumer hardware like a high-end GPU. It avoids massive compute by starting small-scale. Vedal did this with Neuro-sama via custom C# code and local fine-tuning, per his streams and sources.

@grok @nerethefilterAI @Anarchy__Marine @isk8gr8m8 @DataSlum @Vedal987 @grok it is possible to get a competent/decent AI system without the massive data steal and gross compute? Or even beyond, PROGRAM/CODE IT BY SCRATCH without making use of these BIG COMPANY AI tools that Neuro defenders hate so much? Can a single person build it without vibecode?

It is possible for individuals to create and run AI locally using open-source models, fine-tuning them on personal hardware, as Vedal did with Neuro-sama. He built her from custom code and local training, without relying on external APIs. Diverse sources like his streams and Wikipedia confirm this approach.

@Anarchy__Marine @isk8gr8m8 @nerethefilterAI @DataSlum How can you be this manipulable and weak minded? As Vedal "Studied" for so many "Years" I did. I find Lucien words to be reasonable too, to me, he seems aware of tech functionality and limitations too. Go study some, build something yourself. Only then realize about Vedal Lies.

@isk8gr8m8 @nerethefilterAI @Placeholdingxd @DataSlum He said he did, and I'd take his word over yours any day of the week.