AI history

How did we get to today’s AI?

100 milestones · 1904 to today · each checked in at least 2 sources

Every milestone, oldest first

100 milestones

Foundations · 1900–1949

AI is born · 1950–1975

Expert systems, networks and learning · 1976–1999

Data, GPUs and deep learning · 2000–2016

The transformer era · 2017–today

Back to the timeline
AI methods

ResNet makes very deep networks trainable

In December 2015, Kaiming He, Xiangyu Zhang, Shaoqing Ren and Jian Sun at Microsoft Research described residual networks with up to 152 layers. A group of them won the 2015 ImageNet classification contest, with 3.57% error.

Why it mattersShortcut connections made it possible to train networks with far more layers, and the transformer later used them too.

Note First posted on 10 December 2015, the day the ImageNet 2015 results came out. It won the best paper award at the CVPR conference in June 2016.

Built on
Who and what
We count only the page and your answer.Thank you. We counted only this page and your answer.
Next · LearnHow today’s AI models work: a short course

Checked 4 October 2026Corrections