When game developer John Carmack turned toward AI, he asked Ilya Sutskever how to learn the field quickly. Sutskever gave him a curated reading list that he said contained “90% of what matters today.” Carmack worked through it and said AI’s advances began to make sense to him. section 1.3
The list is described as nearly 30 references spanning papers, blog posts, books, and a college course. Its appeal is a path through a crowded field: a small set of ideas selected by a researcher central to contemporary AI. But the original contents remain undisclosed, so the “90%” claim is best understood as Sutskever’s promise, not a verified measure of coverage. section 1.0 section 1.3
Why Carmack’s reaction mattered
Carmack expected difficult, perhaps opaque advances. After reading, he described many as relatively simple techniques that made sense once understood. His account helped turn a private learning aid into a widely discussed guide—despite the uncertainty over its entries. section 1.3
The list also invites a question larger than which references it contains: how do AI’s technical advances connect to competing priorities? Sutskever’s intellectual trajectory brings scaling and alignment, and acceleration and caution, into the same conversation. Interest in that question grew as ChatGPT drew broad attention to AI. section 1.0 section 1.3
