ImageNet

ImageNet appears in 133. 对谢赛宁的7小时马拉松访谈：世界模型、逃出硅谷、AMI Labs、两次拒绝Ilya、杨立昆、李飞飞和42 as the dataset and benchmark context around the AlexNet moment and Fei-Fei Li’s influence. Xie Saining treats it as a problem-definition achievement, not only a data-collection project.

134. 【数据的综述】和谢晨聊，新时代的石油、历史、版图、数据金字塔、定价与Recipe adds 谢晨’s data-industry history. In that source, ImageNet is the early static-dataset stage before Scale AI-style industrial annotation and later Data As Education systems based on feedback, evaluation, and environments.

Why It Matters

The source argues that ImageNet made image classification concrete enough for deep learning progress to become measurable and scalable. This connects the dataset to Problem Definition In Research, Representation Learning, and the later architecture line through ResNeXt and Diffusion Transformers.

Connections

• Fei-Fei Li — figure associated with ImageNet in the source.
• Xie Saining, Kaiming He, and FAIR — researchers and lab context shaped by ImageNet-era vision.
• Representation Learning and Problem Definition In Research — methodological meaning of the dataset.
• Frontier Model Scaling — benchmark/data context for scaling-era AI progress.
• Scale AI, Data As Education, and Data Engine Learning Loop — later data-industry stages added by episode 134.