The Worlds I See

Fei-Fei Li’s memoir, running from her family’s arrival in New Jersey and the years she spent working in her parents’ dry-cleaning shop to building ImageNet — the labelled dataset that gave deep learning something big enough to learn on, and set off the computer-vision decade that followed. It is a research story told as a personal one, and it ends in argument: about what gets built, and who decides.

Context: one of only two non-fiction entries on an otherwise heavily science-fiction list. a16z’s note frames it as reading “like a startup founder’s journey, except the product is modern AI,” and describes the arc as running from “maybe computers can identify edges?” to deepfakes — part immigrant story, part research memoir, part warning.

Connections

• a16z Infra Reading List — the list this came from; Non-fiction section, its own entry.
• Fei-Fei Li — the author and subject.
• Computer Vision — the field ImageNet reorganised.
• Deep Learning — what ImageNet made trainable at scale.
• Machine Learning — the broader discipline.
• Artificial Intelligence — the arc the memoir sits inside.
• a16z — the firm whose infrastructure team curates the list.