Event ·Multi-source

学术研究在大型模型时代面临挑战，经典方法被GPT-6 Astra超越

作者在ECCV 2026前反思学术计算机视觉研究，其论文VIGA是首个用代理方法解决图像到3D场景转换的模型，但被GPT-6 Astra超越。作者认为当前学术工作常基于过时一年甚至两年的想法，导致论文发表后很快过时。CVPR上许多作者仍在研究“旧”问题，而工业界论文虽描述已发布数月的产品，但提供了更完整的分析。

Current read

Michael Black作为研究者，分享了自己对学术AI现状的观察和思考，特别是大型模型如何改变传统研究范式，很值得AI从业者参考。

11 sources58° AI heatLast updated 2026/9/18 02:38:40

Evidence

11 sources

Primary sourceAI Will

2026/9/18 02:38:40View source

Related source 1arXiv: OpenAI

2026/9/17 05:36:18View source

Related source 2Gary Marcus

2026/9/16 04:22:08View source

Related source 3OpenAI Blog

2026/9/16 12:00:00View source

Related source 4Julien Chaumond

2026/9/16 17:59:33View source

Related source 5Justine Moore

2026/9/16 19:01:25View source

Related source 6lmarena.ai

2026/9/16 19:13:56View source

Related source 7掘金本周最热

2026/9/17 04:51:39View source

Related source 8歸藏(guizang.ai)

2026/9/17 08:22:18View source

Related source 9Decoder

2026/9/17 09:45:55View source

Related source 10The Rundown AI

2026/9/17 19:02:20View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story