The Godmother of AI on jobs, robots & why world models are next | Dr. Fei-Fei Li
From Lenny's Podcast
Dr. Fei-Fei Li•Co-Director, Stanford Human-Centered AI Institute & Founder, World Labs
Executive Summary
Fei-Fei Li, a key figure in the AI revolution, discusses her foundational work on the ImageNet dataset, which provided the large-scale data necessary to unlock modern deep learning.
The conversation introduces 'world models' as the next frontier in AI, moving beyond language to spatial intelligence.
Li's new company, World Labs, has launched Marble, the first large-scale generative model for creating interactive 3D worlds.
A core theme is the necessity of a 'human-centered' approach to AI, emphasizing that technology is a tool whose impact is determined by human choices, a principle institutionalized at Stanford's Human-Centered AI Institute (HAI), which Li co-founded.
The episode highlights the dramatic shift in AI's perception over the past decade, from a niche, almost discredited field to a central pillar of the modern tech industry, a transformation Li's work helped catalyze.
8 quotes
Concerns Raised
The potential for AI to be misused if not guided by human-centered principles and robust governance.
A disconnect between Silicon Valley technologists and the broader public, leading to a lack of shared understanding and goals.
The risk of developing AI without sufficient public and cross-disciplinary policy engagement.
Opportunities Identified
The development of 'world models' to unlock spatial intelligence for applications in robotics, design, and scientific discovery.
Using AI as a tool to augment human capabilities and creativity across all professions, from art to healthcare.
Leveraging AI to solve complex societal challenges, such as improving healthcare efficiency and responding to disasters.