
Sign up to save your podcasts
Or
Dr. Gašper Beguš is a UC Berkeley professor of linguistics who studies the interface between human, machine, and animal language. We head into the conversation with a question - is there something fundamentally different about the way that humans learn and the way that machines like LLMs learn? Vector embeddings of the relatedness of language and the map that we carry in our heads of abstract concepts don’t seem that different at the end of the day. This leads us into a discussion of the ways in which humans acquire language, how language evolves, evidence for abstract thought in animals, where the bright line of consciousness can be drawn, and if taking a different approach to training computers to think can generate a machine that can match us in drive and curiosity.
4.6
5050 ratings
Dr. Gašper Beguš is a UC Berkeley professor of linguistics who studies the interface between human, machine, and animal language. We head into the conversation with a question - is there something fundamentally different about the way that humans learn and the way that machines like LLMs learn? Vector embeddings of the relatedness of language and the map that we carry in our heads of abstract concepts don’t seem that different at the end of the day. This leads us into a discussion of the ways in which humans acquire language, how language evolves, evidence for abstract thought in animals, where the bright line of consciousness can be drawn, and if taking a different approach to training computers to think can generate a machine that can match us in drive and curiosity.
428 Listeners
247 Listeners
982 Listeners
572 Listeners
3,411 Listeners
1,242 Listeners
1,049 Listeners
588 Listeners
256 Listeners
1,033 Listeners
258 Listeners
472 Listeners
51 Listeners
130 Listeners
263 Listeners