
Jensen Huang: NVIDIA - The $4 Trillion Company & the AI Revolution | Lex Fridman Podcast #494
Jensen Huang discusses NVIDIA's extreme co-design approach and rack-scale engineering that powers the AI computing revolution
In this episode, Lex Fridman sits down with Luis and Joao Batalha, the co-founders of Fermat's Library, to discuss their mission of making scientific papers more accessible and enjoyable to read. The conversation begins with the brothers sharing their personal backstories and what inspired them to create a platform dedicated to annotating research papers. They explain that many brilliant papers remain hidden behind paywalls and dense academic language, making them inaccessible to curious minds who could benefit from their insights.
Fermat's Library emerged from a simple observation: scientific papers are often written in ways that make them unnecessarily difficult to understand, even for educated readers. The platform addresses this by providing annotated versions of important papers, adding contextual information, explanations of technical concepts, and historical background about the research. This approach transforms how people engage with scientific literature, turning what can be an isolating experience into a collaborative learning process.
The discussion moves into practical advice on how to read and understand research papers effectively. Luis and Joao share their personal techniques for approaching a new paper, including how to scan abstracts, identify key sections, and gradually build understanding rather than expecting to comprehend everything on the first read. They emphasize that reading papers is a skill that improves with practice and that developing good strategies early can significantly enhance one's ability to engage with complex scientific material.
The brothers also address the broader issues within scientific publishing. They highlight how the current system creates unnecessary barriers to knowledge, from expensive journal subscriptions to the pressure on researchers to publish in prestigious venues regardless of the actual quality or impact of their work. This discussion touches on fundamental problems with incentive structures in academia and how they affect the dissemination of knowledge.
A significant portion of the conversation focuses on note-taking and active engagement with papers. The Batalha brothers discuss their personal methods for annotating and tracking important ideas, explaining how the act of writing notes helps cement understanding and creates a personal reference system for future research. They believe that passive reading is less effective than active engagement, where readers question the material, relate it to other knowledge, and create their own summaries.
Throughout the episode, they share some of their favorite papers available on Fermat's Library, giving listeners insight into which research has captured their attention and why. The conversation also includes reflections on using Twitter as a community-building tool, sharing how Fermat's Library has leveraged social media to reach people interested in science and research.
Toward the end, the discussion turns to entrepreneurship and what it takes to build a successful startup. The brothers reflect on their journey, discussing the importance of solving a problem you genuinely care about, understanding your users deeply, and maintaining persistence through challenges. The episode concludes with lighter topics including sports and science fiction, before offering advice for young people interested in research, learning, or building something meaningful.
“The goal of Fermat's Library is to make reading papers not just easier, but actually enjoyable.”
“Most scientific papers are written in a way that makes them unnecessarily hard to understand.”
“Reading papers is a skill that you get better at with practice and good strategies.”
“The scientific publishing system has real structural issues that create barriers for both researchers and the curious public.”
“If you want to build something meaningful, you have to genuinely care about solving the problem and understanding your users.”