PulseAugur
实时 16:53:32
English(EN) Ted Chiang says AI isn't conscious and he's right, but the more useful finding is how mundane the gaps actually are. MIRROR notices a VLM solves a geometry prob

Ted Chiang:AI 的局限性是平凡的,而非缺乏意识

Ted Chiang 认为当前的 AI 没有意识,并指出其局限性令人惊讶地平凡。一个视觉语言模型 (VLM) 在以文本形式呈现时能够解决一个几何问题,但在以图像形式呈现时却无法解决同一个问题。这表明 AI 中的一些“推理失败”可能源于跨模态检索问题,而不是缺乏底层知识。 AI

影响 表明一些 AI 推理失败可能源于跨模态检索问题,而不是根本性的局限性。

排序理由 由一位知名的、有影响力的声音撰写的评论文章,讨论 AI 的能力。

在 Mastodon — sigmoid.social 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

Ted Chiang:AI 的局限性是平凡的,而非缺乏意识

报道来源 [1]

  1. Mastodon — sigmoid.social TIER_1 English(EN) · [email protected] ·

    Ted Chiang says AI isn't conscious and he's right, but the more useful finding is how mundane the gaps actually are. MIRROR notices a VLM solves a geometry prob

    Ted Chiang says AI isn't conscious and he's right, but the more useful finding is how mundane the gaps actually are. MIRROR notices a VLM solves a geometry problem when fed the text but fails the identical problem as an image. The knowledge is in there, one input path just can't …