为什么现在大型语言模型（LLMs）已经变得如此便宜，却还没有语音角色扮演游戏（RPG）呢？
好的，这件事困扰我好几周了。我在手机上玩一些抽卡游戏（别评判），突然意识到……在所有人工智能的热潮和语音技术变得超级便宜的情况下，真正的语音基础角色扮演游戏在哪里呢？
说真的，我们现在有：
- 不再糟糕的语音合成技术（像elevenLabs等）
- 能够在5分钟以上保持角色一致性的语言模型
- WebRTC现在运行得很好
- 手机可以处理语音处理
而我仍然像个原始人一样在点击对话树。最接近的可能是一些人临时拼凑的ChatGPT语音模式，但那根本算不上游戏，对吧？
我明白为什么大型游戏公司做不到这一点——你不能把一个70亿参数的模型和你的游戏一起发布，哈哈。但网页开发者呢？移动开发者呢？只需调用一个API就可以了。
难道只是游戏开发者和人工智能从业者之间没有沟通？即使推理成本降低，是否仍然太高？大家都在等苹果或谷歌先解决这个问题？
老实说，我很惊讶连一个基本的“AI地牢，但你可以和它对话”的东西都没有。AI角色聊天应用已经火爆，显然有需求。
也许我在这里遗漏了什么明显的东西，但这感觉就像是个显而易见的机会。我是不是疯了，还是这个想法就在那里等着有人去实现？
有没有人在实际开发类似的东西？或者知道阻碍是什么？我开始觉得我应该自己做个原型，但我的游戏设计技能……充其量也只是可疑。
查看原文
ok this has been bugging me for weeks. was grinding some gacha trash on my phone (don't judge) and realized... with all the AI hype and voice stuff getting super cheap, where are the actual voice-based RPGs??<p>like seriously, we have:<p>voice synthesis that doesn't suck anymore (elevenLabs and co)
LLMs that can sorta keep characters consistent for more than 5 minutes
webrtc works fine now
phones can handle voice processing
and i'm STILL clicking through dialogue trees like some caveman. closest thing is probably people jerry-rigging chatgpt voice mode but that's not really a game is it.<p>i get why AAA can't do this - you can't ship a 70b model with your game lmao. but web devs?? mobile?? just hit an API and call it a day.<p>is it just that game devs and AI people don't talk to each other? are the costs still too high even with cheaper inference? everyone waiting for apple/google to solve it first?<p>honestly shocked there isn't even a basic "AI dungeon but you talk to it" thing. the AI character chat apps blew up so there's obviously demand.<p>maybe i'm missing something obvious here but this feels like such a gimme. am i crazy or is this just sitting there waiting for someone to build it?<p>anyone actually working on something like this? or know what the blocker is? starting to think i should just prototype something myself but my game design skills are... questionable at best