"How do hallucinations differ in discrete systems, [such as] graphs, text domains than in continuous domains," said @Grigoris_c, at the Simons Institute workshop on Trustworthy AI: From Hallucinations to Reliable Autonomy. "This is exciting for me." Talk: youtube.com/watch?v=Cx47YnPm…
"Diffusion models are used way beyond generating images. They are used to generate diffusion policies for self-driving cars, even protein design. The application area for diffusion models is ever expanding," said @Grigoris_c of @UWMadison during his talk at the Simons Institute.
3/3 For example, OpenAI's Astra 6.1, when prompted to place 31 Stanford tree mascots in front Sather Gate in Berkeley, produced an image with 33 mascots, and a misplaced Campanile behind the gate. "Geographically, it's not possible," said Mahdi Soltanolkotabi of @USC.
2/3 "If you have simple, visual reasoning questions...the models tend to get them wrong," said Mahdi Soltanolkotabi of @USC at the Simons Institute workshop on Trustworthy AI: From Hallucinations to Reliable Autonomy. Video: simons.berkeley.edu/talks/ma…
1/ 3 "The progress in reasoning models has been really crazy recently...But the frontier is quite jagged. Despite all of this progress, there are some very weird shortcomings that still exist, even in the latest models," said Mahdi Soltanolkotabi of @USC at the Simons Institute.
⟳ Simons Institute for the Theory of Computing retweeted
2/2 Their task: to discuss the immediate future of theoretical computer science in light of the challenges posed by AI. Read a report of that closed-door meeting, in the Polynomial Times (see Theoretical Computer Science Meets Its AI Moment, page 3): simons.berkeley.edu/PolyTime…
⟳ Simons Institute for the Theory of Computing retweeted
1/2 The day after OpenAI dropped a bombshell about "a solution to the Navier–Stokes existence and smoothness problem," 32 theoretical computer scientists met for a closed door meeting (that had been planned for months) at the Simons Institute for the Theory of Computing.
Our earlier post should have mentioned that @AnthropicAI announced on the same day that it had independently disproven ETH.
More coverage: science.org/content/article/…
“Immediate, practical applications? There aren’t any,” Dr. Williams said.
“This could lead to better A.I.,” Dr. Impagliazzo said, “in all sorts of ways.”
Theoretical Computer Science Meets Its AI Moment
The day after OpenAI dropped a bombshell about "a solution to the Navier–Stokes existence and
smoothness problem," 32 theoretical computer scientists met for a 2-day closed door meeting (that had been planned for months) at the Simons Institute for the Theory of Computing, to discuss and debate their field's response to AI. In this issue of the @SimonsInstitute's annual magazine, Polynomial Times, is my report of that meeting. Read the article and the entire issue here: simons.berkeley.edu/PolyTime…
Here's my latest on the increasing gaps between open-weight and frontier #AI models revealed in the #1stproof third batch results, announced earlier this week: scieye.wordpress.com/2026/10…