The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) Does ChatGPT “Think”? A Cognitive Neuroscience Perspective with Anna Ivanova - #620
47 snips
Mar 13, 2023 In a thought-provoking discussion, Anna Ivanova, a postdoctoral researcher at MIT Quest for Intelligence, dives into the relationship between large language models and human cognition. She dissects her paper on formal and functional linguistic competence, exploring the cognitive abilities behind language use. Topics include the limitations of AI in achieving true intelligence, the complexities of reporter bias, and the intriguing phenomena of individual differences in perception and cognition. Ivanova's insights challenge conventional notions of AI and intelligence, making for a captivating listen.
AI Snips
Chapters
Books
Transcript
Episode notes
Formal vs. Functional Linguistic Competence
- LLMs excel at formal linguistic competence (grammar, patterns).
- They lack functional competence: applying language to real-world understanding and intent.
Functional Competence and AGI
- Functional linguistic competence involves applying language to real-world tasks, not just knowing the rules.
- It's tied to general intelligence but also includes non-linguistic domains like spatial reasoning.
ChatGPT's Reasoning Abilities
- ChatGPT demonstrates improved paraphrasing and style imitation.
- However, its ability to generate truly novel arguments remains questionable.
