Gasper Begus

If clapping is inappropriate during movements, why does YouTube think it can slap me with commercials during movements?!

Available online now…

Available online now class 2.

YouTube video by Gasper Begus Deep Language Learning LSA Summer Institute Class 2

Deep Language Learning class 2 now online. Learn about neuroscience, interpretability, and decoding non human communication.

youtu.be/oDpmxmlTofM

YouTube video by Gasper Begus Deep Language Learning LSA Summer Institute Class 2

Deep Language Learning class 2 now online. Learn about neuroscience, interpretability, and decoding non human communication.

youtu.be/oDpmxmlTofM

YouTube video by Gasper Begus Deep Language Learning LSA Summer Institute 2025

Our class on Deep Language Learning at the LSA Summer Institute now available online:

www.youtube.com/watch?v=YJnm...

Libraries are monuments to one of the most consequential discoveries in human history that allowed us to transfer ideas in space and time.

I'm pictured here with a dictionary of one of my favorite languages, Vedic Sanskrit, by one of my favorite scholars of all time Grassmann.

Libraries are monuments to one of the most consequential discoveries in human history that allowed us to transfer ideas in space and time.

I'm pictured here with a dictionary of one of my favorite languages, Vedic Sanskrit, by one of my favorite scholars of all time Grassmann.

The Schloss Dagstuhl Institute is awesome! You get sequestered into this beautiful German castle and discuss an important topic at a big picture level.

Punchline: linguistics can offer a lot to the study of LLMs!

Excited for the first class of Deep Language Learning: Modeling Language from Raw Speech at the @lingsocam.bsky.social

YouTube video by Interspecies Internet “New Patterns in Animal Communication with AI: Ethical and Legal Implications" with Gašper Beguš

What implications does finding language-like properties in non-humans have for law, ethics, and philosophy?

In the talk, I outline how AI interpretability can help us make new discoveries in science:

www.youtube.com/watch?v=mzde...

@projectceti.bsky.social

YouTube video by Interspecies Internet “New Patterns in Animal Communication with AI: Ethical and Legal Implications" with Gašper Beguš

What implications does finding language-like properties in non-humans have for law, ethics, and philosophy?

In the talk, I outline how AI interpretability can help us make new discoveries in science:

www.youtube.com/watch?v=mzde...

@projectceti.bsky.social

Discussing LLMs and linguistics in this castle.

Join in a few hours to hear about the legal and ethical implications of AI-driven discovery in animal communication.

“In the Middle Ages, God stood at the top of the value pyramid; the Renaissance elevated humankind to that position. Now we have reached a point once again where the uniqueness—or singularity—of the human being is coming into question.”

In Slovenian newspaper of record:

🐋 TOMORROW! 🚨

Join @begus.bsky.social to explore sperm whale “codas” and what they mean for language, AI, & law.

📅Sat, July 19 | 🕓 4PM GMT / 12PM EST / 9AM PST
📍Online

🔗 RSVP now: https://www.interspecies.io/lectures/gasper-begus

#AIethics #AnimalCommunication #ProjectCETI

“In the Middle Ages, God stood at the top of the value pyramid; the Renaissance elevated humankind to that position. Now we have reached a point once again where the uniqueness—or singularity—of the human being is coming into question.”

In Slovenian newspaper of record:

“In the Middle Ages, God stood at the top of the value pyramid; the Renaissance elevated humankind to that position. Now we have reached a point once again where the uniqueness—or singularity—of the human being is coming into question.”

In Slovenian newspaper of record:

As chatbots improve, humans' unique language abilities are becoming less special UC Berkeley researchers say large language models have gained "metalinguistic ability," a hallmark of human language and cognition no other animal has displayed.

#Anthropology

As chatbots improve, humans' unique language abilities are becoming less special UC Berkeley researchers say large language models have gained "metalinguistic ability," a hallmark of human language and cognition no other animal has displayed.

🤯 LLMs now show “metalinguistic ability”—understanding *about* language itself! A cognitive leap once thought uniquely human. ✨ #AI 🧠

Source: https://phys.org/news/2025-07-chatbots-humans-unique-language-abilities.html

As chatbots improve, humans' unique language abilities are becoming less special UC Berkeley researchers say large language models have gained "metalinguistic ability," a hallmark of human language and cognition no other animal has displayed.

Recent analysis shows advanced AI chatbots are now capable of metalinguistic tasks, such as identifying sentence ambiguity and recursion, narrowing the gap between human and machine language abilities. doi.org/g9rtct

As chatbots get smarter, humans' unique language abilities are becoming less special - Berkeley News UC Berkeley researchers say large language models have gained "metalinguistic ability," a hallmark of human language and cognition no other animal has displayed.

For the first time, metalinguistics and recursion are not exclusively human abilities. Large language models can do them too.

This research is an example of “behavioral interpretability” of LLMs—using prompting to understand inner representations of AI.

news.berkeley.edu/2025/07/14/a...

As chatbots get smarter, humans' unique language abilities are becoming less special - Berkeley News AI platforms like ChatGPT are widely understood to be sophisticated prediction machines. Trained on vast troves of content ranging from news articles and books to film scripts and Reddit posts, they anticipate the next most likely letters and words when...

As chatbots get smarter, humans' unique language abilities are becoming less special - Berkeley News
->UC Berkeley | #Chatbot #AI | More info from EcoSearch

As chatbots get smarter, humans' unique language abilities are becoming less special - Berkeley News UC Berkeley researchers say large language models have gained "metalinguistic ability," a hallmark of human language and cognition no other animal has displayed.

UC Berkeley researchers say large language models have gained "metalinguistic ability," a hallmark of human language and cognition no other animal has displayed. news.berkeley.edu/2025/07/14/a...

As chatbots get smarter, humans' unique language abilities are becoming less special - Berkeley News UC Berkeley researchers say large language models have gained "metalinguistic ability," a hallmark of human language and cognition no other animal has displayed.

For the first time, metalinguistics and recursion are not exclusively human abilities. Large language models can do them too.

This research is an example of “behavioral interpretability” of LLMs—using prompting to understand inner representations of AI.

news.berkeley.edu/2025/07/14/a...

Would you find this useful?

bsky.app/profile/begu...

The phonology of sperm whale coda vowels https://www.biorxiv.org/content/10.1101/2025.06.09.658556v1

Pleased to see this published (open access, of course). I have a paper on Celtic historical phonology in this volume, which you can get here doi.org/10.5281/zeno.... 1/🧵