Watch and track your favorite playlist.
Curated by: Ontology Explained: Philosophy and AI (109 videos)
This is the third video in this playlist: https://www.youtube.com/playlist?list=PLSjZcAM1g_Us Are AI models "grown" like crops, or engineered like software? And why are AI safety writers so obsessed with painting LLMs as bizarre, unpredictable alien minds? In this video, we continue our chapter-by-chapter breakdown of "If Anyone Builds This, Everyone Dies" by analyzing Chapter 2. We put their provocative "Grown, Not Crafted" analogy to the test, untangle the messy dialectic around LLM autocomplete vs. "true understanding," and address the infamous incident where Bing Chat allegedly threatened a philosophy professor. While there is a modest, valid point buried here—that modern deep-learning models are too complex for humans to predict in every scenario—the broader claims about AI preferences and "alien minds" are woefully under-defended and rely on subtle leaps in logic. RESOURCES & REFERENCES: • Chauhan & Geiger (2024) Paper: https://openreview.net/pdf?id=6gvM1koUTl • Sonakshi & Geiger Poster: https://neurips.cc/media/PosterPDFs/NeurIPS%202024/104102.png?t=1732923618.844852 • Agassi on Becker's Tongue Tell: https://www.youtube.com/watch?v=ja6HeLB3kwY • ELIZA Effect: https://en.wikipedia.org/wiki/ELIZA_effect • Pareidolia: https://en.wikipedia.org/wiki/Pareidolia • Adam Dunn vs. Tony Gwynn Breakdown: https://bleacherreport.com/articles/400786-adam-dunn-better-than-hall-of-famer-tony-gwynn If you enjoy charitable philosophical analysis of AI risk literature, make sure to like, subscribe, and share! Leave your thoughts in the comments below—I read every single one.