Lex Fridman Podcast #74 – Michael I. Jordan: Machine Learning, Recommender Systems, and the Future of AI
10 snips
Feb 24, 2020 Michael I. Jordan, a renowned professor at Berkeley, is a pivotal figure in machine learning and AI. In this engaging conversation, he examines how far we truly are in AI development, discussing the misperceptions surrounding technology's progress. He critiques recommender systems and their impact on consumer trust, particularly in the context of Facebook's privacy issues. Additionally, Jordan explores the profound distinctions between engineering feats and scientific breakthroughs while advocating for a human-centric approach to artificial intelligence.
AI Snips
Chapters
Transcript
Episode notes
AI as an Emerging Engineering Discipline
- Current AI developments resemble the emergence of chemical or electrical engineering from fundamental sciences.
- Building valuable, human-centered systems using data and decisions is key, but the engineering aspect is still ad hoc.
Cluelessness About Brain Computation
- We are clueless about how the brain performs computation, making current AI metaphors about brain inspiration misleading.
- Neuroscience will require centuries of research to understand the brain's complexity.
Misconceptions About Brain-Inspired AI
- We lack a fundamental understanding of how thought emerges from the brain's complex processes.
- Claims of brain-inspired AI breakthroughs are often overstated and misleading to young researchers.