Happy Chanukah, everybody! There is a lot of academic activity around, and the ceasefire in Gaza has brought some relief and hope. Let me tell you about the (unusually high number of) lectures I attended in November 2025, in reverse chronological order. I have known many of the speakers for several decades—together they represent centuries of acquaintance.

I would like to highlight the lecture by Orit Raz at the HUJI combinatorics seminar, where she presented a mind-boggling program aimed at improving the bounds for the unit distance problem using rigidity theory of frameworks. I will start with the lecture by John Martinis, the 2025 Nobel Prize laureate, and end with a lecture by Steve LaValle from Finland on robotics.

The full line of speakers was: Steve LaValle (Nov. 3), Oz Almog (Nov. 6), Shmuel Weinberger (Nov. 11), Yuri Gurevich (Nov.13), Micha Sharir (Nov. 16), Meir Feder (Nov. 19), Itai Benjamini (Nov. 20), Amichai Lampert (Nov. 20), Orit Raz (Nov. 24), Adi Shamir (Nov. 24), Dorit Aharonov, Erez Berg, Shai Evra, and Snir Gazit (Nov. 25), Ron Wettenstein (Nov 27), Shakhar Smorodinsky (Nov. 30), Roy Meshulam (Nov. 30), Yaron Oz, Yonatan Cohen, and John Martinis (Nov. 30).

November’s lectures

Sunday November 30: Physics Nobel lecture

John Martinis, the 2025 Nobel Laurates in Physics, talked about his famous 1985 experiment on quantum tunneling and about superconducting quantum computers. He also answered many questions.

According to John Martinis, the current pace of progress will not yield useful quantum computers in the coming decades (before “most of the members of the audience will retire”). He outlined some ideas and plans for an alternative trajectory. (Click to enlarge the picture.)

Additional short talks were given by Yaron Oz and Yonatan Cohen. On December 1, Martinis gave a similar talk at HUJI.

I first met John Martinis at our 2013 QSTART conference, where he gave a lecture about quantum error correction based on superconducting qubits (video). Yonatan Cohen was our host in the recent visit to the IQCC. Yaron Oz is a well-known high-energy physicist.

Sunday November 30: Two combinatorics lectures

At the TAU combinatorics seminar, Shakhar Smorodinsky explained his work with Noga Alon, which we discussed in this post. Roy Meshulam talked (over Zoom) about Cayley complexes, codes, and homology at the Bar Ilan combinatorics seminar.

Roy Meshulam is one of my closest collaborators. We started collaborating in the 1980s and wrote our first paper together in the early 2000s. Our joint work is mainly on topological Helly-type theorems. I have known Shakhar Smorodinsky for at least two decades, and he spent a year as my postdoc in Jerusalem.

Thursday, November 27: Explainability in AI (Reichman University)

Ron Wettenstein lectured in Reichman’s CS colloquium on: “From Game Theory and Boolean Logic to Decision Tree Explainability”. Ron described his work with Alex Nadel on WOODELF: a unified and efficient algorithm for computing Shapley values on decision trees.

Ron is a PhD student at Reichman University under the supervision of Udi Boker, Alex Nadel and me.

Tuesday November 25: The theory of quantum computing (HUJI, IIAS)

This was the opening day of a new interdisciplinary center for the theory of quantum computing. The speakers were Dorit Aharonov, Erez Berg, Shai Evra, and Snir Gazit. It is almost 15 years since the kick-off of the quantum science center at HUJI, and since then similar centers have opened at other Israeli universities. Recently, four or five national centers were established.

Dorit presented a list of central problems on the agenda of the new center, many of which are related to quantum error-correcting codes. Shai talked about the mathematics of quantum error-correcting codes; Snir gave a statistical-mechanics view of surface codes and other physical states; and Erez discussed methods of error mitigation and related experiments on IBM quantum computers.

Dorit Aharonov has been my colleague for the past three decades. (I already count her as a colleague during her doctoral years, but she was indeed a mature scientist even then.) A few years earlier, she was a student in my “advanced calculus” class. Shai is a mathematician at HUJI, and I have known him for more than a decade. I met Erez and Snir at the conference itself. (Both were students of the late Israeli physicist Assa Auerbach whom I mentioned in this post.)

Monday November 24: The unit distance problem and rigidity

Orit Raz gave a talk at the HUJI combinatorics seminar. Can graph (framework) rigidity be used to push down the Trotter–Szemerédi bounds for unit distances in the plane? This is a fascinating research direction. What is needed, among other things, is an understanding of rigidity for non-generic embeddings, which is an important question in its own right. Will the approach described by Raz to the unit distance problem lead to a success similar to the Elekes approach to the distinct distance problem? Time will tell.

Here are links to the relevant papers:

1. Erdős’s unit distance problem and rigidity, János Pach, Orit E. Raz, József Solymosi (2025)
2. Dense graphs have rigid parts, Orit E. Raz, József Solymosi (2019)
3. Configurations of lines in space and combinatorial rigidity, Orit E. Raz (2016)

Orit was my colleague at HUJI, and she recently moved to Ben Gurion University. Let me mention a series of startling works by Orit Raz with Alan Lew, Eran Nevo and Yuvel Peled around rigidity.

Monday, November 24: Cryptography and neural networks

At the HUJI CS colloquium, Adi Shamir talked about joint work with David Gerault, Anna Hambitzer and Eyal Ronen. They show that natural implementations of block ciphers (such as DNNs) on neural networks can be broken in linear time using non-Boolean inputs, and they develop a new practical method for implementing any desired cryptographic functionality as a standard ReLU-based DNN in a provably secure and correct way.

Cryptography for neural networks. (Where is Adi?)

Adi Shamir is probably the most famous cryptographer in the world. Cryptography represents a gap in my understanding of theoretical computer science (I complained about it here in 2018), although Alon Rosen has contributed a great deal to changing the situation. I probably first met Adi in the mid-1980s.

Thursday, Nov 20, Probability with a little combinatorics

At the TAU probability seminar, Itai Benjamini described some terrific problems (and some solutions) in probability theory. The title was “Cayley Graphs vs. Random Graphs.” There is a lot of material here for this blog—stay tuned.

Itai Benjamini has been a close collaborator for many decades. Even earlier, in 1987, he was a student in my “convexity” class—one of the best classes I ever taught.

Shortly after Itai’s talk, Amichai Lampert gave an impressive lecture in the TAU number theory seminar on number theory and dynamics.

Wednesday, November 19: Information theory and neural networks

Meir Feder gave a talk at the TAU learning seminar about an information-theoretic framework for understanding modern machine learning. Meir described how his work in information theory since the 1990s is relevant to the neural-network revolution.

The new results are described in the paper:

Meir Feder, Ruediger Urbanke, Yaniv Fogel, Information-Theoretic Framework for Understanding Modern Machine-Learning

Meir Feder and I were together at MIT in the early 1980s, and our paths have crossed many times since. Meir is an Oscar Laurate (see this post).

SSunday, November 16: Computational Geometry — Per Aspera Ad Astra (TAU CS Colloquium) (video)

Micha Sharir described his scientific journey from early papers on piano movers through the early days of computational geometry until today.

I have known Micha Sharir personally since the mid-1970s (and by name since the late 1960s). How did I already know about Micha as a teenager, years before I first met him? The answer appears in this interview.

Thursday, November 13: Abstract State Machines (Reichman University)

Speaker: Yuri Gurevich

Title: What’s an algorithm?

Abstract: The classical/basic notion of algorithm was elucidated in the1930s–1950s. Starting from the 1960s, this notion has been expanded to probabilistic algorithms, quantum algorithms, etc. In the 1980s the speaker introduced abstract state machines (ASMs), and in 2000 he axiomatized basic algorithms and proved that every basic algorithm is step-for-step simulated by an appropriate basic ASM. The axiomatization has served both theoretical purposes (notably, proving the original Church-Turing thesis) and for practical purposes (notably, enabling the development of an ASM-based tool that Microsoft’s Windows Division used to produce numerous high-level executable specifications required by the EU). In the talk we define an elegant (at least in our view) generalization of basic algorithms: basic interactive algorithms, which may interact with human and artificial partners. It turns out that probabilistic and quantum are naturally basic interactive. We axiomatize basic interactive algorithms and prove that every such algorithm can be step-for-step simulated by a basic interactive ASM — opening the door to new applications.

This was a fascinating talk about abstract state machines—a powerful theoretical and applied tool in computer science—and about Yuri Gurevich’s remarkable path from mathematics to logic, theoretical computer science, applied computer science, and even quantum computation. Yuri told us that his mother was the dominant person at home, and that when the Germans were approaching their town during World War II, it was a rare case in which his father insisted that the family move east; this decision saved their lives.

I attended an impressive lecture on average-case complexity (and the theory of Leonid Levin) that Yuri Gurevich gave at IBM Almaden in 1991. (I probably also encountered Yuri in Israel in the 1970s.) We met at Microsoft Research and have been friends since the late 1990s.

Yuri at Reichman University’s sculpture garden

Wednesday, November 12: Morse complexity of homological classes (TAU)

Shmuel Weinberger talked about “How Complex Are the Fundamental Structures Underlying Manifolds?”

Here is the description in TAU’s Institute for advanced Studies page:

In his lecture on Morse complexity of homology classes, Prof Weinberger will discuss a refined approach to understanding the topology of manifolds. Building on Gromov’s 1970s pseudonorm and ideas inspired by Thurston, he will explore the concept of minimizing the number of critical points in a Morse function for a manifold representing a homology class. While this aligns with Gromov’s approach in dimension two, higher dimensions reveal striking differences.

The lecture will touch on connections to classical topology—including open book decompositions and surgery theory—representation theory, and elliptic operators, highlighting joint work with Manin and Tshishiku.

Shmuel Weinberger is an eminent geometer and topologist and he is interested in application as well, particularly in “persistent homology“. Here is a post featuring an article he wrote “about conjectures“. I think we first met in the 90s.

Thursday, November 5: Reichman University — The End of the Era of the Academic Degree

Oz Almog, a sociologist from Haifa University talked about the end of the Era of the Academic Degree. A lot of food for thought in this provocative talk. It reflects the academic research of Oz Almog and his wife Tamar Almog who wrote together a book “Academia all the Lies”.

TAU CS Colloquium, November 2: Fundamental Challenges in Robotics and Embodied AI

Steve LaValle, Robotics (video)

Here is the abstract to Steve’s lecture. “The field of robotics is wildly exciting and rapidly gaining worldwide attention, yet it is often an enigma in terms of its scope and scientific foundations. Throughout the decades, it has been variously viewed as an application field of more mature disciplines such as computer science (AI, algorithms, machine learning) and mechanical engineering (kinematics, dynamics, nonlinear control).

In this Computer Science Colloquium lecture, Professor LaValle will argue that robotics has its own unique and growing scientific core, with deep questions and modeling challenges that should inspire new directions in computer science, engineering, and mathematics.

Let me mention that, unlike in other areas where AI (and deep learning) has dominated the scene, the situation in robotics is very different, and it is unclear what role AI will ultimately play.

Steve is very well known, but since we belong to different communities, I met him for the first time at this lecture. Steve was impressed by the honesty and modesty of the Finnish people and decided to make Finland his home.

Steve’s praise for Finland reminded me of the opening ceremony of the ICM 2022, where the President of Finland offered greetings. There were no trumpets when he entered, and the audience was not asked to stand. The President gave a five-minute welcoming speech, complimented mathematicians for their efforts and contributions, and concluded by shyly inviting us to enjoy the summer weather in Helsinki. Then he waved his hand and left.

November lectures — a collage of speakers

There were quite a few other November talks that I missed (and in a few cases I caught up privately with the speakers). All in all, it looks like the levels of academic activity and enthusiasm have returned to those before the war. Naturally, however, the number of foreign visitors is still considerably lower.