I've recently been getting very into GPT-3, a predictive text AI. It's fun to run my groupchats through it and see how it simulates conversation. For example, this is a conversation between me and my friend Joanna, generated entirely by bot:
Joanna: this is a very good podcast
Adam: I like the different segments
Joanna: I like that they are highlighting the variation and diversity of the araucanian language
Adam: I feel like you have a very good understanding of the araucanians.
Joanna: gotta keep it
Adam: Do you think there were other tribes who spoke it
Joanna: is there an araucanian word for when a bird is chirping?
Adam: I think there were different tribes who spoke it
Joanna: i think i will do some research on this
Joanna: because it's a really cool language and I want to learn more about it
Joanna: and I want to learn more about the people who speak it
Adam: like who
Joanna: the araucanian people
Adam: the araucanians are a people?
Joanna: they are an indigenous group in what is now chile and argentina
Joanna: yeah, they're a people
Adam: do you know what language they spoke
Joanna: that i know of, they spoke araucanian
It's crazy how accurately this mimicks our speech and text patterns. And this tool has so much potential!
Hello! I'm Adam Aleksic. This year, I graduated from Harvard University with a degree in Government and Linguistics. There, I co-founded the Harvard Undergraduate Linguistics Society and wrote a thesis on Serbo-Croatian language policy, magna cum laude. In addition to etymology, I also really enjoy philosophy, trivia, vexillology, geography, board games, conlanging, art history, and law.