The world's leading venue for collaborative research in theoretical computer science. Follow us at YouTube.com/SimonsInstitute.
Berkeley, CA
- Tweets 1,492
- Following 351
- Followers 10,054
- Likes 820
Simons Institute for the Theory of Computing retweeted
We all know that AI is a highly divisive issue. A working group convened at the Simons Institute earlier this month, to formulate concrete action items for TCS academia to take in response to AI. Remarkably, despite the wide range of opinions, participants were able to arrive at broad consensus on the list of recommendations here: simons.berkeley.edu/AITCSrep…. Many apply more broadly to mathematics, and beyond.
This document is NOT a declaration or essay. It is a list of concrete proposals, fleshed out all the way from motivation to implementation. Its purpose is to provide a common reference point and shared baseline for the community to discuss and enact change.
The actions are intentionally designated to be taken in the short term (specifically, this academic year); they are not meant to represent a stable vision of the future. You do not have to like them; you just have to like them MORE than what will happen if nobody takes action. Not everybody at the workshop agreed with every action, but they ultimately came together through discussion and compromise.
For transparency, I was involved in this event in a purely organizational capacity. I did not come up with any of these actions, but I recognize their importance and I will support them however I can. I hope you do too.
In September, we hosted a working group to think about how the TCS community should respond to AI progress. Their report, released yesterday, highlights 12 near-term actions that received broad support among the participants.
simons.berkeley.edu/news-pub…
2
13
50
13,505
Simons Institute for the Theory of Computing retweeted
Here are some of my main takeaways from the exciting #Cryptography reunion workshop I attended as a Science Communicator in Residence @SimonsInstitute this summer -
scieye.wordpress.com/2026/09…
Many thanks to the organizers, Elette Boyle, Yuval Ishai, Abhishek Jain & Daniel Wichs!
1
3
388
In September, we hosted a working group to think about how the TCS community should respond to AI progress. Their report, released yesterday, highlights 12 near-term actions that received broad support among the participants.
simons.berkeley.edu/news-pub…
1
21
60
21,277
1/4 Are diffusion language models ready for the real world? Not quite. Today's diffusion language models are missing some key ingredients, said @volokuleshov of @Cornell, at the Simons Institute workshop on Diffusion Generative Modeling: Progress and Next Steps
1
2
9
1,411
3/4 Block diffusion addresses this limitation. "You'd run diffusion over blocks of tokens in parallel and generate the following block of tokens, using a fixed length diffusion block, conditioned on the [generated] tokens," said @volokuleshov of @Cornell at the Simons Institute.
1
338
4/4 "This allows you to interpolate between auto-regression and diffusion," said @volokuleshov of @Cornell at the Simons Institute workshop on Diffusion Generative Modeling: Progress and Next Steps. Video: simons.berkeley.edu/talks/vo…
219
Join us Tuesday, 9/29 for the first Richard M. Karp Distinguished Lecture of this academic year, featuring @BooleanAnalysis.
simons.berkeley.edu/events/m…
7
29
2,361
We invite your project proposals for Circles, the Simons Institute – Jane Street Small Group Collaborations, which supports groups of 3–6 researchers for 4 weeklong gatherings over 2 years.
Apply by Oct. 15 (deadline extended).
simons.berkeley.edu/particip…
2
15
97
15,497
Simons Institute for the Theory of Computing retweeted
1/5 The case for diffusion language models: "A lot of the [early] gains in language modeling performance have come from scaling pre-training...[the training algorithm] was designed to be very parallelizable across GPUs," said @volokuleshov of @Cornell at the Simons Institute
1
2
17
3,679
1/5 The case for diffusion language models: "A lot of the [early] gains in language modeling performance have come from scaling pre-training...[the training algorithm] was designed to be very parallelizable across GPUs," said @volokuleshov of @Cornell at the Simons Institute
1
2
17
3,679
4/5 "To carry [over the pre-training gains] to post-training and inference time scaling...we need to develop language models that are fully parallel both in training and inference," said @volokuleshov of @Cornell at the Simons Institute. Diffusion language models are an option.
1
1
2
373
5/5 Volodymyr Kuleshov, @volokuleshov, of @Cornell spoke at the Simons Institute workshop on Diffusion Generative Modeling: Progress and Next Steps. Video: simons.berkeley.edu/talks/vo….
1
4
346
Simons Institute for the Theory of Computing retweeted
1/5 Imagine taking an fMRI scan of the brain of a person viewing an image, and reconstructing what the person saw from the fMRI alone. That's Brain-IT. "This is state-of-the-art image decoding from fMRI," said Michal Irani (@WeizmannScience), at the Simons Institute.
1
3
18
2,091
1/5 Imagine taking an fMRI scan of the brain of a person viewing an image, and reconstructing what the person saw from the fMRI alone. That's Brain-IT. "This is state-of-the-art image decoding from fMRI," said Michal Irani (@WeizmannScience), at the Simons Institute.
1
3
18
2,091
4/5 Here are some examples of successful reconstructions of the images observed by a person, using only the fMRI images. For each pair, left is the original image, right is the reconstruction from the fMRI scan. Paper: arxiv.org/abs/2510.25976 Roman Beliy et al.
1
1
1
279
"Of course, we also have failures," said Michal Irani, at the Simons Institute workshop on Topics in Intelligence: World Models and Social Reasoning. For e.g., the fMRI scan of a person viewing a cat became a bear. Video: simons.berkeley.edu/talks/mi…
Paper: arxiv.org/abs/2510.25976
1
269