Simons Institute for the Theory of Computing
@simonsinstitute.bsky.social
The world's leading venue for collaborative research in theoretical computer science. Follow us at http://YouTube.com/SimonsInstitute.
Thurs., 2/5 at 3:30 p.m., Law & Society Fellows Rui-Jie Yew and Greg Demirchyan discuss Alignment Problems in AI Governance.
simons.berkeley.edu/events/align...
simons.berkeley.edu/events/align...
Alignment Problems in AI Governance
In an event comprising short talks and dialogue, Simons Institute Law and Society Fellows Rui-Jie Yew and Greg Demirchyan will explore two challenges of alignment in AI governance. First, we currently...
simons.berkeley.edu
January 31, 2026 at 12:55 AM
Thurs., 2/5 at 3:30 p.m., Law & Society Fellows Rui-Jie Yew and Greg Demirchyan discuss Alignment Problems in AI Governance.
simons.berkeley.edu/events/align...
simons.berkeley.edu/events/align...
January 29, 2026 at 8:46 PM
1/2 "The most striking finding we had is that the students that practiced math problems with ChatGPT without any guardrails did 17% worse on immediate subsequent exam where they did not have AI assistance." Hamsa Bastani of @upenn.edu at the Simons Institute. simons.berkeley.edu/talks/hamsa-...
January 26, 2026 at 4:20 PM
1/2 "The most striking finding we had is that the students that practiced math problems with ChatGPT without any guardrails did 17% worse on immediate subsequent exam where they did not have AI assistance." Hamsa Bastani of @upenn.edu at the Simons Institute. simons.berkeley.edu/talks/hamsa-...
1/2 Given two randomized control trials, RCT1 (100 participants) & RCT2 (1000), intuition says that RCT2 gives the better estimate of treatment effect. But sometimes our intuition is wrong, said Hannah Li of @columbiauniversity.bsky.social at the Simons Institute. simons.berkeley.edu/talks/hannah...
January 24, 2026 at 3:24 AM
1/2 Given two randomized control trials, RCT1 (100 participants) & RCT2 (1000), intuition says that RCT2 gives the better estimate of treatment effect. But sometimes our intuition is wrong, said Hannah Li of @columbiauniversity.bsky.social at the Simons Institute. simons.berkeley.edu/talks/hannah...
Join us next week for the Federated and Collaborative Learning Boot Camp. Register to attend in person or access the livestream.
simons.berkeley.edu/workshops/fe...
simons.berkeley.edu/workshops/fe...
January 24, 2026 at 12:05 AM
Join us next week for the Federated and Collaborative Learning Boot Camp. Register to attend in person or access the livestream.
simons.berkeley.edu/workshops/fe...
simons.berkeley.edu/workshops/fe...
Register to attend or access the livestream for Katrina Ligett's Karp Lecture on 1/29:
simons.berkeley.edu/events/lets-...
simons.berkeley.edu/events/lets-...
January 23, 2026 at 11:57 PM
Register to attend or access the livestream for Katrina Ligett's Karp Lecture on 1/29:
simons.berkeley.edu/events/lets-...
simons.berkeley.edu/events/lets-...
In this Richard M. Karp Distinguished Lecture in the Complexity and Linear Algebra program, Virginia Vassilevska Williams examines progress on matrix multiplication algorithms over the decades and offers some intuition about where the research area may be headed.
simons.berkeley.edu/news/matrix-...
simons.berkeley.edu/news/matrix-...
On Matrix Multiplication Algorithms | Richard M. Karp Distinguished Lecture
Fast matrix multiplication is a central goal in algorithms research. The goal is to find the smallest real value omega such that n by n matrices can be multiplied in n{omega + o(1)} time in the worst ...
simons.berkeley.edu
January 23, 2026 at 9:52 PM
In this Richard M. Karp Distinguished Lecture in the Complexity and Linear Algebra program, Virginia Vassilevska Williams examines progress on matrix multiplication algorithms over the decades and offers some intuition about where the research area may be headed.
simons.berkeley.edu/news/matrix-...
simons.berkeley.edu/news/matrix-...
The successes of generative AI and LLMs involve deep internal representations of the world. In his Theoretically Speaking lecture, Jon Kleinberg explored how these work, and how they relate to the representations of the world that we build as humans.
bit.ly/4qE98dO
bit.ly/4qE98dO
January 22, 2026 at 1:15 AM
The successes of generative AI and LLMs involve deep internal representations of the world. In his Theoretically Speaking lecture, Jon Kleinberg explored how these work, and how they relate to the representations of the world that we build as humans.
bit.ly/4qE98dO
bit.ly/4qE98dO
Our workshop on LLMs, Cognitive Science, Linguistics, and Neuroscience explored why LLM capabilities in logic and reasoning lag behind their linguistic capacity, and how a different model — the human brain — could point the way forward.
bit.ly/45QWwaM
bit.ly/45QWwaM
Uncovering the Role of Language in Machine Reasoning
Large language models (LLMs) gain their encyclopedic knowledge and conversational tact by learning from an entire internet’s worth of human-generated text. But learning from language alone has shown d...
bit.ly
January 20, 2026 at 8:27 PM
Our workshop on LLMs, Cognitive Science, Linguistics, and Neuroscience explored why LLM capabilities in logic and reasoning lag behind their linguistic capacity, and how a different model — the human brain — could point the way forward.
bit.ly/45QWwaM
bit.ly/45QWwaM
Congratulations to Sushant Sachdeva on winning the Infosys Prize! And the Infosys video on his work includes some fun footage of Calvin Lab...
youtu.be/39tFMYUCI6c?...
youtu.be/39tFMYUCI6c?...
The sparsest graph of them all - Sushant Sachdeva
YouTube video by Infosys Prize
youtu.be
January 20, 2026 at 8:17 PM
Congratulations to Sushant Sachdeva on winning the Infosys Prize! And the Infosys video on his work includes some fun footage of Calvin Lab...
youtu.be/39tFMYUCI6c?...
youtu.be/39tFMYUCI6c?...
Reposted by Simons Institute for the Theory of Computing
In my latest blog, I discuss the challenges of building a powerful #cryptographic technique that aims to “obfuscate” the internal implementation details of programs- lnkd.in/gT3smPZh
Thanks Seyoon Ragavan and Rahul llango for the nice chat about their works @simonsinstitute.bsky.social!
#TCS
Thanks Seyoon Ragavan and Rahul llango for the nice chat about their works @simonsinstitute.bsky.social!
#TCS
Constructing indistinguishability obfuscation schemes in a hardness rich world
Rachel Lin’s talk, “Expedition to Obfustopia,” at the Simons Institute examined indistinguishability obfuscation (IO) in cryptography. She highlighted the challenges in constructi…
lnkd.in
January 16, 2026 at 8:57 PM
In my latest blog, I discuss the challenges of building a powerful #cryptographic technique that aims to “obfuscate” the internal implementation details of programs- lnkd.in/gT3smPZh
Thanks Seyoon Ragavan and Rahul llango for the nice chat about their works @simonsinstitute.bsky.social!
#TCS
Thanks Seyoon Ragavan and Rahul llango for the nice chat about their works @simonsinstitute.bsky.social!
#TCS
In the latest In this episode of our Polylogues web series, Simons Institute Founding Associate Director Alistair Sinclair interviews newly appointed Institute Director Venkatesan Guruswami.
www.youtube.com/watch?v=VY37...
www.youtube.com/watch?v=VY37...
Venkatesan Guruswami | Polylogues
YouTube video by Simons Institute for the Theory of Computing
www.youtube.com
January 17, 2026 at 2:22 AM
In the latest In this episode of our Polylogues web series, Simons Institute Founding Associate Director Alistair Sinclair interviews newly appointed Institute Director Venkatesan Guruswami.
www.youtube.com/watch?v=VY37...
www.youtube.com/watch?v=VY37...
1/2 An oldie — in LLM years — but goodie! Ilya Sutskever speaks on his observations about a theory of unsupervised learning at the Simons Institute workshop on Large Language Models and Transformers, in Aug 2023. Video: simons.berkeley.edu/talks/ilya-s...
January 14, 2026 at 7:26 AM
1/2 An oldie — in LLM years — but goodie! Ilya Sutskever speaks on his observations about a theory of unsupervised learning at the Simons Institute workshop on Large Language Models and Transformers, in Aug 2023. Video: simons.berkeley.edu/talks/ilya-s...
1/4 Do LLMs understand? "They understand in a way that’s very different from how humans understand," Dileep George, @dileeplearning.bsky.social, of Google DeepMind at the Simons Institute workshop on The Future of Language Models and Transformers. Video: simons.berkeley.edu/talks/dileep...
December 27, 2025 at 1:53 PM
1/4 Do LLMs understand? "They understand in a way that’s very different from how humans understand," Dileep George, @dileeplearning.bsky.social, of Google DeepMind at the Simons Institute workshop on The Future of Language Models and Transformers. Video: simons.berkeley.edu/talks/dileep...
1/2 "LLMs are amazing, but hard to control. Diffusion models are also amazing, and (maybe too) easy to control." @cornelluniversity.bsky.social's Kilian Weinberger, at the Simons Institute, on how these contrasting properties can be used to steer LLMs. Video: simons.berkeley.edu/talks/kilian...
December 24, 2025 at 4:47 AM
1/2 "LLMs are amazing, but hard to control. Diffusion models are also amazing, and (maybe too) easy to control." @cornelluniversity.bsky.social's Kilian Weinberger, at the Simons Institute, on how these contrasting properties can be used to steer LLMs. Video: simons.berkeley.edu/talks/kilian...
1/4 LLMs learn language competency, factual common knowledge and factual tail knowledge. Factual knowledge has a heavy tail and should be stored in a database. “It doesn’t belong in the model weights”: Cornell's Kilian Weinberger at the Simons Institute. Video: simons.berkeley.edu/talks/kilian...
December 21, 2025 at 4:11 AM
1/4 LLMs learn language competency, factual common knowledge and factual tail knowledge. Factual knowledge has a heavy tail and should be stored in a database. “It doesn’t belong in the model weights”: Cornell's Kilian Weinberger at the Simons Institute. Video: simons.berkeley.edu/talks/kilian...
1/2 "In the LLM scenario...jailbreaking is very powerful. In agentic scenarios, it gets even more dangerous." @sivareddyg.bsky.social of @mila-quebec.bsky.social at the Simons Institute's workshop on Safety-Guaranteed LLMs Video: simons.berkeley.edu/talks/siva-r...
December 15, 2025 at 2:28 PM
1/2 "In the LLM scenario...jailbreaking is very powerful. In agentic scenarios, it gets even more dangerous." @sivareddyg.bsky.social of @mila-quebec.bsky.social at the Simons Institute's workshop on Safety-Guaranteed LLMs Video: simons.berkeley.edu/talks/siva-r...
1/2 "You could have bad goals and you could be smart, but if you can’t do anything in the world, then you can’t do a lot of harm. The trio is the thing that kills us." @yoshuabengio.bsky.social, at his Richard M. Karp Distinguished Lecture at the Simons Institute: www.youtube.com/watch?v=g0lj...
Superintelligent Agents Pose Catastrophic Risks — ... | Richard M. Karp Distinguished Lecture
YouTube video by Simons Institute for the Theory of Computing
www.youtube.com
December 11, 2025 at 6:23 AM
1/2 "You could have bad goals and you could be smart, but if you can’t do anything in the world, then you can’t do a lot of harm. The trio is the thing that kills us." @yoshuabengio.bsky.social, at his Richard M. Karp Distinguished Lecture at the Simons Institute: www.youtube.com/watch?v=g0lj...
1/2 "Symmetries play a fundamental role in machine learning." Soledad Villar of @jhu.edu speaking at the Simons Institute workshop on Randomness, Invariants, and Complexity. Video: simons.berkeley.edu/talks/soleda...
December 9, 2025 at 6:00 AM
1/2 "Symmetries play a fundamental role in machine learning." Soledad Villar of @jhu.edu speaking at the Simons Institute workshop on Randomness, Invariants, and Complexity. Video: simons.berkeley.edu/talks/soleda...
Join us next Tuesday, 12/9 for Jon Kleinberg's talk in our Theoretically Speaking public lecture series. Registration is required.
simons.berkeley.edu/events/ais-m...
simons.berkeley.edu/events/ais-m...
AI’s Models of the World, and Ours | Theoretically Speaking
When we see someone performing at the top of their craft, we often marvel at both their observable achievements and the hidden internal expertise that they’ve accumulated. Something similar is true, i...
simons.berkeley.edu
December 4, 2025 at 6:00 AM
Join us next Tuesday, 12/9 for Jon Kleinberg's talk in our Theoretically Speaking public lecture series. Registration is required.
simons.berkeley.edu/events/ais-m...
simons.berkeley.edu/events/ais-m...
1/3 "What are the conditions for an AI system in the future to cause catastrophic harm?" Turing Award winner @yoshuabengio.bsky.social asked, during his Richard M. Karp Distinguished Lecture at the Simons Institute earlier this year. www.youtube.com/watch?v=g0lj...
Superintelligent Agents Pose Catastrophic Risks — ... | Richard M. Karp Distinguished Lecture
YouTube video by Simons Institute for the Theory of Computing
www.youtube.com
November 29, 2025 at 4:55 AM
1/3 "What are the conditions for an AI system in the future to cause catastrophic harm?" Turing Award winner @yoshuabengio.bsky.social asked, during his Richard M. Karp Distinguished Lecture at the Simons Institute earlier this year. www.youtube.com/watch?v=g0lj...
1/2 Networks are creating bottlenecks for AI. There's "a separation between the growth of computation for [AI] and the growth of networking infrastructure...We must make our networks more efficient." Chen Avin of @bengurionuni.bsky.social at the Simons Institute. simons.berkeley.edu/talks/chen-a...
November 28, 2025 at 4:09 AM
1/2 Networks are creating bottlenecks for AI. There's "a separation between the growth of computation for [AI] and the growth of networking infrastructure...We must make our networks more efficient." Chen Avin of @bengurionuni.bsky.social at the Simons Institute. simons.berkeley.edu/talks/chen-a...
Join us for Jon Kleinberg's Theoretically Speaking lecture, 12/9! This is an in-person event, and registration is required.
simons.berkeley.edu/events/ais-m...
simons.berkeley.edu/events/ais-m...
November 27, 2025 at 2:08 AM
Join us for Jon Kleinberg's Theoretically Speaking lecture, 12/9! This is an in-person event, and registration is required.
simons.berkeley.edu/events/ais-m...
simons.berkeley.edu/events/ais-m...
1/2 "What are graph foundation models? How are they different from language models? Or maybe not," asked Mikhail Galkin of @googleresearch.bsky.social at the Simons Institute's workshop on Graph Learning Meets Theoretical Computer Science. Video: simons.berkeley.edu/talks/mikhai...
November 16, 2025 at 11:48 AM
1/2 "What are graph foundation models? How are they different from language models? Or maybe not," asked Mikhail Galkin of @googleresearch.bsky.social at the Simons Institute's workshop on Graph Learning Meets Theoretical Computer Science. Video: simons.berkeley.edu/talks/mikhai...
We were delighted to host Simons Institute Industry Day 2025 today, full of lively exchanges and talks by partner companies, program organizers, and postdoctoral-level research fellows.
November 14, 2025 at 3:09 AM
We were delighted to host Simons Institute Industry Day 2025 today, full of lively exchanges and talks by partner companies, program organizers, and postdoctoral-level research fellows.