This """free game""" is a transparent attempt by squeenix to extract & harvest free language data from players to help train their algorithmic-AI tech. I can't tell you what to do with your gaming time, but it seems like a real bad deal to provide free labor to a (increasingly shady and unethical) megacorp and help them train the same AI models that are gonna be used to exploit and demoralize you. This game is one of the great early visual novels, and very much worth playing. Thankfully there's already a...

i dont mind ai as long the game work

geforce gtx 960? 15 gb of room? what? why?

Can't believe they're already implementing this tech into video games. Really excited to see fully immersive NPCs with unlimited conversation options! Can you imagine what else they'll come up with?

Dealing with the language model, I can see a few bugs with the English implementation that are making things more finicky than they should be. The technique makes sense and is fairly standard Data Science, with Sentence Embeddings and a Sentence Similarity calculation, but there are a few specific problems eat into the "error budget". These may be exclusive to some of the English implementations. - Maximum Similarity is not 100%: An exact string match looks like it only has a 97 or 98% similarity or ...