The Dissenter

#1119 Anna Ivanova: Language and Large Language Models

12 snips
Jul 4, 2025
Join Anna Ivanova, an Assistant Professor at Georgia Tech, as she explores the intricate connection between language and cognition. Discover how advanced brain imaging techniques like fMRI reveal language processing in the brain. Ivanova discusses the capabilities and limitations of large language models, questioning whether they truly understand language like humans do. She also compares programming languages with natural languages, shedding light on how they activate different brain networks. This conversation dives deep into the fascinating world of language and AI!
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

Language Network Is Specialized

  • Language activates a distinct brain network separate from vision and audition.
  • It processes language specifically, not general abstract compositional tasks like math or music.
INSIGHT

Language Network Not for Categorization

  • Language areas are not recruited for non-linguistic tasks like feature-based categorization.
  • People can categorize objects without engaging the brain's language network.
INSIGHT

Distributed But Stable Language Network

  • The brain uses a distributed but stable network for language processing.
  • Broca's and Wernicke's areas form part of this language network, but it includes other regions too.
Get the Snipd Podcast app to discover more snips from this episode
Get the app