[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/vt/ - Virtual Youtubers

Search:


View post   

>> No.46912183 [View]
File: 707 KB, 1204x1101, 1659871081477861.png [View same] [iqdb] [saucenao] [google]
46912183

>>46911279
>announces model reveal stream
>reveals it before stream
Somebody post that dum gif.

>> No.46667426 [View]
File: 707 KB, 1204x1101, 1659871081477861.png [View same] [iqdb] [saucenao] [google]
46667426

>>46666134
Last thread someone suggested Holocure and I support the idea. I think gauntlet-styled games in general and bullet hells could be good for Neuro since they're usually simple enough in gameplay and can be fun to watch. Touhou is the obvious candidate here.
Remembering Miyune collab, I think it can be interesting to have some cooperative game and have other players try to keep Neuro alive, like an escort mission of sorts.
And in remembrance of Mario Retardy streams, maybe party games could work here, for both collab and gameplay material. Cart games as well.

>>46666672
Leeching is the easiest way to gain traction in streaming. There are plenty of streamers who became popular purely by riding on someone else's success.

>>46666891
The only problem is that you'd have to individually make AI for each game. I think it's better to focus on one game that will last longer, at least for now.

>> No.46059021 [View]
File: 707 KB, 1204x1101, 1659871081477861.png [View same] [iqdb] [saucenao] [google]
46059021

>>46057561
From what little we know about him, he's a pretty young coder and probably never had much of an exposure to public or dealing with business before, and now there's suddenly thousads of people watching his creation and expecting shit of him and expecting it now. I'd say he holds up decently for someone his age, though it does feel like he's getting ahead of himself now.

>>46057577
The difference now is that we're getting closer to the AI that could make decisions on its own, decisions that might affect humans. Before that all we had was basically superpowered calculators that did precisely what was requested of them.
From what I understand, the trick of machine learning is that you don't need to teach it everything by yourself, you just feed it the data and hope it'll give you the desired result. You might have the general idea about how it processes the data and why it gives certain responces, but there isn't really a proper way to debug or foolproof it. You get what you get.
No matter how good the algorithm is, there's always a chance of it not getting it right. Would you want to trust your life to such a thing?

Navigation
View posts[+24][+48][+96]