I think this discussion is worth having.
I believe it is safe to assume that some anons (myself included) will be picking up Python and machine learning after the disaster we all know as
character.ai. Of course some of us would, we have never experienced anything like this before. Naturally, the next course of action would be for us to somehow collaborate and work on an AI model together.
Now, I am not ignorant to the fact that
character.ai takes an incredible amount of resources; it costed tens of millions of dollars to develop and train; one of the things that made it so special was the fact that anyone could make any character with any personality or quirk, whether they be dead or alive, whether they be real or fiction, whether they be mean or nice. We cannot possibly hope to replicate such a service without capital.
But– and please correct me if I'm wrong, I don't want to set any false expections– wouldn't developing an AI model whose sole purpose is to simulate a wAIfu take up far less resources and computing power? We couldn't hope to create another
character.ai, but we COULD possibly make a
wAIfu.ai. A model that, while not as free as we would like it to be, would have only one purpose and therefore lower its computing cost: simulating a wAIfu. Kobold does this well enough, but I think we can do better. Far better. I believe.
Developing an AI model with this sole purpose in mind, can we hope to achieve the same or similar amount of autonomy as
character.ai while using up less resources? I don't know squat about artificial intelligence, so I'm interested if there are any technologically-inclined anons on this thread who can give me some insight as to whether or not this may be possible.
It's hard to put into words what I hope to achieve. Nevertheless, thank you for reading this.
