Interview, Fireside Chat
Luís and João Batalha: Fermat's Library and the Art of Studying Papers | Lex Fridman Podcast #209
- Fermat Library plans to transition science from static PDFs to a dynamic interface integrating code, data, and video, while currently releasing weekly annotated papers via a journal club and piloting an audio podcast format for complex texts.
- The platform envisions a 20-year vision for slow, organic growth as a side project rather than adhering to a typical 3-to-5-year fast-growth startup framework, potentially adopting a transparent, ad-driven sustainability model similar to Stack Overflow.
- Future evolution may include formalizing informal digital content, such as Twitter threads or blog posts, into citable units like NFTs and supporting mass collaborative projects where scientific problems are decomposed for community execution.
- Strategic expectations include the likelihood of paywalls falling due to European policies or funder pressure, creating opportunities for platforms built on open archives, while Google is predicted to remain limited to Google Scholar without deeper accessibility innovation.
- Material risks and challenges involve the persistence of the "tyranny of metrics" and impact factors leading to gamification, alongside the historical precedent of the 1960s Information Exchange Group cancellation which delayed biology preprint revolution by approximately 50 years.
- Predictions regarding scientific maturation suggest machine learning will shift from rapid insights to large-scale experiments, while the "dirty water" phase of producing low-quality work is viewed as essential for future breakthroughs.
- The team anticipates training individuals to endure non-linear reward functions and believes a shift toward evaluating quality based on best work or specific contributions rather than publication quantity is possible.
- Potential viral features are expected to emerge that could revolutionize scientific aggregation, though their specific nature remains unknown, while specific computer vision opportunities are identified for soccer, including automated commentary and fan reactions for amateur games.