两篇Mastodon帖子认为，大型语言模型不具备智能，而是作为预测性过滤器运行。作者认为，大型语言模型似乎发明的“语言”仅仅是放大的噪声或反馈啸叫，而非真正的交流。他们声称，人类也不是用语言思考的，并以棒球运动员王贞治为例，认为他通过进入一种“无意识”状态而非口头思考而取得了成功。 AI
影响 这些帖子挑战了AI智能的概念，认为模型是复杂的预测性过滤器，而非有意识的实体。
排序理由 该集群包含来自社交媒体平台的观点文章，讨论AI模型的性质。
在 Mastodon — mastodon.social 阅读 →
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
完整方法见我们的编辑标准。
How do I know they are very stupid? Because imagine hitting a baseball if you had to think in words. Imagine Sadaharu Oh hitting all those home runs, if he had to think in words. In fact he did it by deliberately going into a ‘Zen’ state of ‘mindlessness’. He practiced this with …
This was just a bunch of filters amplifying each other’s noise! The didn’t ‘invent their own languages’. That was the LLM token version of feedback squeal. You must not think of things as ‘intelligences’. They are predictive filters. Human intelligence does not work that way. The…