
Have you ever wondered what happens when you look inside the “mind” of artificial intelligence? Or why and how we mishear song lyrics? In this episode, Computer Science and Linguistics double major Aidan Mokalla breaks down their ambitious dual-thesis research, exploring how large language models internally represent semantic content across languages and how noise alters human language perception.
Give this episode a listen to learn more about completing two distinct theses at Reed (an uncommon feat) and hear about Aidan’s trajectory from prospective intellectual property lawyer to NYU research fellow. Aidan also offers practical advice on managing thesis anxiety and shifting research questions, while emphasizing the importance of the student–advisor relationship and making the case for why understanding AI’s internal processes is critical to creating more trustworthy, equitable technology.
Reed community members can read Aidan’s Computer Science thesis, “Language-Independent Concept Spaces Found in Large Language Models,” and Linguistics thesis, “Modulated Signal Filtering: Wouldn’t it be Noise?” online in the Electronic Theses Archive.
