Letter from the Director, January 2025

Dear friends,

Greetings from Berkeley, where the spring semester is just getting underway. This term, we are hosting the second half of the Special Year on Large Language Models and Transformers. This program explores LLMs from a broad perspective, asking questions about their power, scalability, trustworthiness, safety, implications for research in the sciences, and the legal frameworks in which LLMs operate, as well as their impacts on society at large. Researchers from computer science, mathematics, neuroscience, cognitive science, linguistics, law, and other disciplines continue on at the Simons Institute this spring. 

Meanwhile, we are preparing to host a joint workshop with SLMath on AI for Mathematics and Theoretical Computer Science. Scheduled to take place April 7 to 11 in Calvin Lab and SLMath’s Eisenbud Auditorium, this workshop will introduce mathematicians and theoretical computer scientists to emerging technologies for advancing and accelerating mathematical discovery, including automated reasoning, proof assistants, and machine learning. I hope to see many of you there.

In this issue of the newsletter, we present the latest episode of our Polylogues web series, in which former Simons Institute Science Communicator in Residence Ben Brubaker of Quanta Magazine sits down with Boaz Barak, a theoretical computer scientist and cryptographer at Harvard, whose recent move into deep learning research has brought him in-house at OpenAI.

Also in our SimonsTV corner this month, we have two of our favorite talks from the tail end of our fall programs: neuroscientist Bruno Olshausen (UC Berkeley) on the neural computations that underlie animals’ abilities to create and manipulate geometric representations of the world, and a debate on the motion “Current LLM scaling methodology is sufficient to generate new proof techniques needed to resolve major open mathematical conjectures such as P ≠ NP,” featuring Sébastien Bubeck (OpenAI), Tom McCoy (Yale), Pavel Izmailov (Anthropic), and Ankur Moitra (MIT), and moderated by Anil Ananthaswamy (science communicator in residence).

In conclusion, I want to thank all of you who donated to our Annual Fund in 2024! Sustaining high levels of activity at the Institute depends on the generosity of members of our community. Your support means the world to us.

And to all, I hope 2025 is off to a good start for you. We look forward to sharing updates from the Institute over the course of the year.

Best wishes,
Venkat

Venkatesan Guruswami
Interim Director, Simons Institute for the Theory of Computing