TLexDR
Subscribe

Melanie Mitchell: Concepts, Analogies, Common Sense & Future of AI

12-28-19 with Melanie Mitchell ▶ 1h 52m 📖 3 min read
Core Takeaways
Melanie Mitchell critiques the term 'artificial intelligence' as misleading, preferring 'complex information processing.' ▶ 1:00
Why it matters This critique suggests that the term 'AI' may mislead public and academic understanding of machine capabilities.
Deep learning lacks the ability to prioritize relevant features, limiting its understanding of concepts like a paddle or a ball in games. ▶ 45:00
Why it matters This limitation indicates why AI struggles with tasks requiring a deeper understanding, like transferring skills across contexts.
The long tail problem in autonomous driving highlights the challenge of unexpected edge cases not covered in training data. ▶ 1:20:00
Why it matters Addressing the long tail problem is critical for the safe deployment of autonomous vehicles in real-world conditions.
Mitchell argues that existential threats from AI are distant, while immediate threats like nuclear weapons are more pressing. ▶ 1:40:00
Why it matters Focusing on immediate threats could redirect resources from speculative AI risks to more urgent global issues.
Concepts and analogies are crucial for cognition, yet current AI struggles to form and use them fluidly. ▶ 30:00
Why it matters Without mastering concepts and analogies, AI cannot achieve human-like reasoning or common sense.

How the conversation moved

Lex Fridman opens the conversation by questioning the adequacy of the term 'artificial intelligence,' which Melanie Mitchell critiques as misleading. She suggests 'complex…

Ask this episode Deep

A preview of how Deep chat answers, grounded in this episode with citations and timestamps:

Cite this episode

For papers, blog posts, anywhere.

Copied!

Related episodes

Where to go next from this conversation.

More on these ideas
Jitendra Malik: Computer Vision Shares deep learning, autonomous driving 1h 41m Daniel Kahneman: Thinking Fast and Slow, Deep Learning, and AI Shares artificial intelligence 1h 18m Noam Chomsky: Language, Cognition, and Deep Learning Shares deep learning, cognition 35m Wojciech Zaremba: OpenAI Codex, GPT-3, Robotics, and the Future of AI Shares deep learning 2h 51m

AI-generated summary · last refreshed 2026-06-08 17:14:35 · how we make these

Quotes are matched verbatim against the source transcript; references are checked to resolve to real URLs. Even so, AI can misread structure or attribute claims imperfectly. If you spot an error, please let us know.

Report an inaccuracy →