When legendary game programmer John Carmack decided to move into AI, he asked OpenAI co-founder Ilya Sutskever what he should read. Ilya handed him a list of around thirty works and said that learning them would cover 90% of what matters in modern deep learning.
The version circulated today was reconstructed by the community. It is a coherent tour from convolutional and recurrent nets, through attention and Transformers, to scaling laws and the information-theoretic roots of learning.
Sources: community reading list, Aman’s AI Journal, and Ilya’s List.