Gašper Beguš on AI
2 quotes from Gašper Beguš on the wall, 2025, newest first. Each one is verbatim and links to where it was said. Sources include Quanta Magazine, University of California.
It appears that we're less unique than we previously thought we were
Linguist Gašper Beguš, commenting on his study showing an AI model performing sophisticated linguistic analysis once thought uniquely human, in Quanta Magazine, 2025.
What it refers to
A study found an AI model performing linguistic analysis tasks, like recursion and ambiguity resolution, once thought to be exclusively human abilities.
Other voices
- We're less special than we thoughtGeoffrey HintonResearcher30 August 2025
- They live in the land of correlations between images and text, not in the land of ideasGary MarcusResearcher31 August 2025
- These models somehow just generalize dramatically worse than peopleIlya SutskeverBuilder27 November 2025
- Large language models are about mimicking peopleRichard SuttonResearcher26 September 2025
- Statistics and machine learning in action, with nothing mystical behind itTeppo FelinResearcher21 November 2025
Quanta MagazineIn a First, AI Models Analyze Language As Well As a Human Expert
In these models, we have one of the rare things that we thought was human-only
Gašper Beguš, UC Berkeley linguist, commenting on research showing AI chatbots can perform metalinguistic analysis once thought uniquely human, July 2025.
What it refers to
Metalinguistic reflection and recursive sentence structuring, abilities Noam Chomsky described as uniquely human language traits.
Other voices
- We're less special than we thoughtGeoffrey HintonResearcher30 August 2025
- They live in the land of correlations between images and text, not in the land of ideasGary MarcusResearcher31 August 2025
- Large language models are about mimicking peopleRichard SuttonResearcher26 September 2025
- Most interesting problems scale extremely badlyYann LeCunResearcher27 April 2025
- Seemingly conscious AIMustafa SuleymanBuilder19 August 2025