General Vtuber Discussion (V1)

[El Rrata]

Duck chooses memes over cute.

 

[BioBreak]

Spoiler: Talent freedom post

Pink cat bad accidentally showed her face on cam
View attachment 6761

Hilarious.

 

[Trap Supremacy]

Spoiler: Talent freedom post

Pink cat bad accidentally showed her face on cam
View attachment 6761

I remember back when I had a brief affair with Nyanners back in 2015 she used to do facecam streams. I know nowadays she wouldn't dare, but having seen her before, this feels like a dumb joke (especially since, unlike the most corpo vtubers, she literally hasn't changed the account she streams from).

 

[kestraline]

Tenma did it way better.

I was just about to post that.



Of course, Cooksie the clipfag wouldn't even know that, because he only cares about vshitshow and their related lewdtubers.

 

[EliteDoragon]

Mikomashi Suinye is back on her radio show

She talks about her post-surgery recovery. Apparently she was so bored that she went to Disney 3 times and had to hum a lot.



Audio:

 

[angry sushi]

Super purge
have the timestamp when Kiara gets raided, she talks about that she likes her voice and tells her to apply for Hololive
Kiara was watching and further encouraging her after she finished her own stream

---

So what you're asking of us is to return back to Orange Woman Bad around here? If so that will take us to 1k pages by the new year easily.

 

[masterstocking]

I appreciate how people just need to insert that fucking monkey into everything.

 

[Shuba]

Mikomashi Suinye is back on her radio show
View attachment 6764

She talks about her post-surgery recovery. Apparently she was so bored that she went to Disney 3 times and had to hum a lot.



Audio:

Nice, I didn't know someone was fully translating Heikousen Scramble. I've only seen clips around but was curious about the whole thing cause Suisei and Koroazu seem like a great pair.


Like here where Suisei realizes she's old and decaying:

 

[ChilChakSack]

Looks like offbrand Pippa made it into this month's YOTC art.

 

[Audi Vide Tace]

I'll take shitposting schizos over no-fun allowed bunch of self-important losers.

The difference here is one statement is asserting that the reddit community itself are closer to being normal functional people, and the other is asserting that the crazy people in /vt are more entertaining to watch.

Look at it this way: which one would you rather have over for dinner with the family? Self-important losers are at least self-regulated, the schizos may stab your aunt when reaching too far for the salt shaker.

 

[Svarog]

View attachment 6762

Holy shit, she is never return

 

[kestraline]

How long until Ina gets a page that says how long she hasn't streamed for, like Ayame had?

 

[thhrang]

View attachment 6766

Looks like offbrand Pippa made it into this month's YOTC art.

Artist confirms.

 

[Hff201]

EDIT:
Karen's released a teaser vid for her upcoming Sinderella album:


Bit of a departure from her previous releases as there doesn't seem to be a central theme and the sound just seems all over the place. There seems to be more prevalent/blatant autotune this time around.

The theme's the seven deadly sins (plus some extras). The more pop-driven sound might've been influenced by the record label, though in fairness Karen's been experimenting beyond rap outside of Hololive too - she did a rock cover recently that I like:

 

[masterstocking]

clock goes chop chop

 

[chihirogumi]

I have a newfound respect for rpr. Chii-chan is well-known for having dedicated antis who streamsnipe and harass her to a deranged degree. It even happened during her solo Master run in Apex (she still achieved it though so lol). Apparently he has been getting their accounts banned behind the scenes and I've noticed the streams lately have been streamsnipe free, so pretty based.

 

[The Proctor]

Chii-chan is well-known for having dedicated antis who streamsnipe and harass her to a deranged degree.

What an earth did she do to manage that?

 

[lakyus]

Spider drawing!

---

Is it too late to take me apple picking? :3c

By Amiya Aranha

 

[chihirogumi]

What an earth did she do to manage that?

It's a long story that dates back to about two years ago. I'll try to give a short summary off the top of my head.

1. Chii-chan get addicted to Apex. She plays so much she decides to enter Apex tournaments.
2. She struggles in the practice games of the first tournament and backseat Apex fans shit up her chat. They also dislike bomb her videos. She doesn't take the heat well and goes on a break.
3. fast forward a couple of months and a popular JP Apex streamer is in a tournament with Chii-chan, he and his fanbase dislike vtubers. She kills him in one of the games and as a result more antis flood her chat with negativity and dislike bombs her videos.
4. The hate brigade gets so big that they shittalk her on twitter and forums, create a discord server dedicated to streamsniping her, and shit up other vtuber and streamers chats by falseflagging as her fans ("you're not as good as chii-chan") for example.
   
   Her worst crime is letting the hate get to her by responding to it on stream, it gets to the point where she starts to cry. They will of course feast on this and push the hate even further in a vicious cycle. Hence why she had that lawsuit filed earlier this year because some schizos were taking it way too far.

 

[Audi Vide Tace]

Eh, not exactly. Trademarking has its own formal process, but copyright is a thing that just exists by default. If you make it, you own it. The exception to this is if you're producing work for someone else and they have you sign a work-for-hire contract, which transfers that default copyright over to them. You can safely assume that any (reputable) corpo owns all their character designs, and indie designs will belong to either the artist or the talent, depending on the arrangement between the two parties.

That's why the AI art situation is a bit murky legally to start with - to the best of my knowledge, there's no court precedent for it to establish exactly how transformative an AI-generated piece needs to be before it's considered legally distinct from the source pieces it pulls from. If it turns out someone's AI artwork is noticeably similar to an existing piece, the original artist might try to take action against it. They don't need to have registered anything anywhere, they can just say "hey, this looks like mine" and fire out DMCAs, starting a legal process.

True, but this exists in regular art too. There are court cases where the defendant was accused of using copyrighted work and being completely unaware of the previous work it was supposedly a derivative of. The rules of what makes something transformative can be a bit of a blurred line at times.

AI-generated piece needs to be before it's considered legally distinct from the source pieces it pulls from

I am sure I am being bit of a pedant here, so let me apologize in advance if you are already aware of this; but, just to make sure everyone is on the same page: the AI isn't "pulling" from the source pieces, it is a neural network that builds associations to entities/concepts during the training phase. This is necessary for computer to actually understand and perceive what that entity is. A computer can't perceive (or think consciously), so how do you teach it what a mountain is?

The source images do not need to be accessible by the AI after the training phase, and most of the time they aren't. If you were to download stable diffusion you wouldn't be downloading the 2.3 billion images it was trained on. Likewise, NovelAI is a custom model of the stable diffusion AI (which is opensource), so it doesn't have access to the original 2.3 billion images it was first developed on either.

Now whether or not an AI is able to understand what makes something "transformative" and apply that concept is a different matter entirely. It certainly has been producing transformative works, but the efficacy rate of the false negative case occurring is still TBD.


Edit: There weren't too many great sources previously that explained in a concise manner on how these newer AI's (specifically stable diffusion and thus, NovelAI) fully functioned, but I just found this recent explanation that gives good insight into stable diffusion. Now you shouldn't need to really piece-chunk information together from half-fleshed out articles; assuming anyone here is interested.