iiinova — where your digital self keeps evolving

Juspreet Singh Sandhu @freegaussians.bsky.social · 21h

Please read. The traditional “unspoken” (but highly human-centered) precedents of the community are under attack.

Reposted by Juspreet Singh Sandhu

Zack Kilpatrick @zpkilpat83.bsky.social · 08/10/2026

We're hiring: tenure-track Assistant Professor in computational or physical applied mathematics at CU Boulder, starting Aug 2027. Nonlinear waves, dynamical systems, applied PDEs, numerics. ML/quantum ties a plus. Full consideration by Nov 8. jobs.colorado.edu/jobs/JobDeta...

jobs.colorado.edu

Assistant Professor

045

Juspreet Singh Sandhu @freegaussians.bsky.social · 07/10/2026

Updates: 1) The pre-print for the SK Parisi measure at SK now explicitly mentions Lopatto and Chen as having proved the same conclusion. 2) The PM ansatz issue I have is marketing -- It's a "final step" not a breakthrough; the breakthrough was Panchenko's original paper.

000

Juspreet Singh Sandhu @freegaussians.bsky.social · 07/10/2026

Quick follow up -- Not only had @oldheneel.bsky.social, @jnshi.bsky.social and I sampled SK all the way to beta < 1 in TVD before oAI, oAI also claimed 1) a proof of fRSB support for the SK's Parisi measure (which was done before), and 2) A proof of the PM ansatz (done in almost all but one regime)

111

Juspreet Singh Sandhu @freegaussians.bsky.social · 07/10/2026

It is a bit interesting to read this. Fortunately, @oldheneel.bsky.social , @jnshi.bsky.social and I were able to sample SK all the way to beta < 1 a few weeks before oAI released 300 pages analyzing Glauber at beta <= 1. The proof is unintelligible and a chunk uses free probability w/o citation

241

Reposted by Juspreet Singh Sandhu

Marius Lemm @mariuslemm.bsky.social · 07/10/2026

I think we need to digest these proofs and then come up with new and harder problems. It’s both disorienting and liberating.

021

Reposted by Juspreet Singh Sandhu

Zach Weinersmith @zachweinersmith.bsky.social · 06/10/2026

OpenAI just blooped hundreds of proofs and counterexamples to open problems: github.com/openai/math/... Boggling.

github.com

math/preprints at main · openai/math

Contribute to openai/math development by creating an account on GitHub.

1822439

Juspreet Singh Sandhu @freegaussians.bsky.social · 06/10/2026

This is the only paper (from July) - arxiv.org/pdf/2605.17338 where the author has comments from known experts and the content skill/complexity/area demonstrates readability and a different expertise than the generalized TAP estimates. The author has written 5 papers in September alone.

arxiv.org

200

Juspreet Singh Sandhu @freegaussians.bsky.social · 06/10/2026

Update - Google scholar for the author: scholar.google.com/citations?hl... Not a single mathematical paper before August 2026. All prior papers are in AI or numerical/theoretic physics. The person is no longer a researcher and has blog posts on LLMs.

scholar.google.com

100

Juspreet Singh Sandhu @freegaussians.bsky.social · 06/10/2026

An interesting paper though this seems like a lot of slip, and on a quick reading I do not believe the author is the architect of most of the proofs. It seems the paper is AI generated in huge chunks; horrible exposition and some incorrect references to parts of our previous paper.

200

Reposted by Juspreet Singh Sandhu

Clément Canonne @ccanonne.github.io · 05/10/2026

This point regarding AI-produced proofs is of course not new, but is beautifully put here. www.quantamagazine.org/is-ai-the-en... @quantamagazine.org

It’s as if you were teleported to the peak of a tall mountain. Surrounded by fog, you have no idea where you are, or what’s around you. You do not know how your mountain connects to others, and you have no equipment to help you explore, no way to help someone else join you. If you had climbed the mountain yourself, you would have experienced how the human body adapts to altitude and changes in oxygen levels. You might have had to invent tools to navigate, to climb steep cliffs, or to make a shelter. You might have encountered a fellow explorer, gotten lost together in a hidden valley, and found a plant that could be turned into a life-saving medicine.

Instead you’re perched on the peak but in the dark, while the maker of the teleportation machine tells you that it can explore the wilderness better than any human.

510924

Juspreet Singh Sandhu @freegaussians.bsky.social · 28/09/2026

The same situation as with Tristan Buckmaster and NS, brought to you now in Biology with Anthropic: www.nytimes.com/2026/09/27/s... In this substack, a scenario leading to this blatant academic plagiarism, which Mark Chen of oAI responded to in Tristan's case with "all companies do it" is given:

nytimes.com

Did Anthropic’s A.I. Really Make a Scientific Discovery on Its Own? (Gift Article)

An expert at the University of Copenhagen said his team had been sharing its research with the company’s A.I. model, Claude, and that its new finding matched their work.

121

Juspreet Singh Sandhu @freegaussians.bsky.social · 28/09/2026

This was an up-and-down journey full of a lot of learning and discovery. However, the embarrassment of missing the fact that we could’ve gone for the “local” regularity properties straight away on top of the free-probability and cavity interpolation theory is real.

010

Reposted by Juspreet Singh Sandhu

Clément Canonne @ccanonne.github.io · 27/09/2026

When I say "this is not good" let me be more specific: (1) this is not good for academia (destroying the current system before it has a chance to reassess its incentive structure), (2) it is not good for science (flooding the gates with ill-understood results and insightless or badly written proofs)

1255

Juspreet Singh Sandhu @freegaussians.bsky.social · 24/09/2026

Finally! Submitted PHA-4 to arXiv: drive.google.com/file/d/1m3gc... Sampling SK all the way to beta < 1 building on the free-probability and cavity interpolation theory developed in PHA-3, with the amazing @oldheneel.bsky.social and @jnshi.bsky.social

drive.google.com

141

Reposted by Juspreet Singh Sandhu

arXiv math.PR Probability @mathpr-bot.bsky.social · 24/09/2026

Omar Abdelghani, Roland Bauerschmidt, Thierry Bodineau, Benoit Dagallier: Log-Sobolev inequality for the sinh-Gordon model arxiv.org/abs/2609.28091 arxiv.org/pdf/2609.28091 arxiv.org/html/2609.28091

021

Reposted by Juspreet Singh Sandhu

Simons Institute for the Theory of Computing @simonsinstitute.bsky.social · 23/09/2026

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/participate/...

033

Reposted by Juspreet Singh Sandhu

Johan Ugander @jugander.bsky.social · 20/09/2026

My Yale colleague Dan Spielman has started a substack, which I thought I'd signal boost since Dan is an outstanding thinker and communicator (and person!) but isn't on social media much. Here's his first substantive post from a month ago: mathadjacent.com/p/llms-testi...

mathadjacent.com

LLMs, Testing, Grade inflation, and Accommodations

Like every professor, I’ve been worrying about how AI systems will force us to change our teaching practices.

1228

Reposted by Juspreet Singh Sandhu

angela zhou @angelamczhou.bsky.social · 16/09/2026

Another blog post: separatedhyperplane.substack.com/p/we-cant-le... We can't let frontier labs choose their own auditors - or their own questions. Triangulating between the near-past lens of algorithmic accountability, the financial crisis, and the AI evaluation landscape as it looks now:

separatedhyperplane.substack.com

We can’t let frontier labs choose their own auditors - or their own questions.

We should start with mandatory audits that adequately cover harms under current laws and regulations while we figure out the rest.

2249

Juspreet Singh Sandhu @freegaussians.bsky.social · 16/09/2026

Short and crisp: a little breath of air in an otherwise (at least somewhat) asphyxiating environment.

010

Reposted by Juspreet Singh Sandhu

arxiv math.MP @arxiv-math-mp.bsky.social · 15/09/2026

Shirshendu Ganguly, Wencai Liu Sharp decay thresholds for eigenvalues of discrete Schr\"odinger operators on $\mathbb{Z}^2$ arxiv.org/abs/2609.15908

011

Reposted by Juspreet Singh Sandhu

arxiv math.MP @arxiv-math-mp.bsky.social · 15/09/2026

Shin-itiro Goto Liouville integrability of Beltrami-Maxwell fields with sources: A geometric theory of magnetic surfaces arxiv.org/abs/2609.14242

021

Reposted by Juspreet Singh Sandhu

Ryan O'Donnell @booleananalysis.bsky.social · 14/09/2026

Outstanding progress towards the Unique Games Conjecture posted by Yumou Fei, Dor Minzer, and Shuo Wang: eccc.weizmann.ac.il/report/2026/... 👀

eccc.weizmann.ac.il

ECCC - TR26-179

3237

Juspreet Singh Sandhu @freegaussians.bsky.social · 13/09/2026

Speaking of real math -- We will have a nice PHA installment (PHA-4) on arXiv in the next week or two to finish the SK story! Stay tuned! PS: If I was focused we could've done this sooner but I got into proving infinte-dim LSIs via Malliavin calculus. cc: @jnshi.bsky.social @oldheneel.bsky.social

021

Juspreet Singh Sandhu @freegaussians.bsky.social · 11/09/2026

Full disclosure: I have signed this. Recent events about lack of attribution/academic honesty, the lack of transparent audits into the technology as no prompts/transcripts were released, and my own experiences for the last year have led me to this.

000

Juspreet Singh Sandhu @freegaussians.bsky.social · 11/09/2026

Sticking to some good-old math: arxiv.org/abs/2609.08169 This came out of the PHA series of papers as @jtnshi & I observed one could Guerra interpolate the planted-SK with SL field to CWRF. We were curious to develop a sharp understanding of the overlaps under the Gibbs at high-temp for the CWRF.

arxiv.org

On overlap concentration in the Curie-Weiss Random Field model

We show that the overlaps of two independent replicas drawn from the Gibbs measure of the Curie--Weiss model with random field (CWRF) exhibit sub-Gaussian tails when the inverse temperature $ β< 1$ an...

242

Reposted by Juspreet Singh Sandhu

Ryan O'Donnell @booleananalysis.bsky.social · 10/09/2026

Am teaching grad complexity theory at CMU; about 1/3 of the lectures will be new (vs. last time), 'modern' results. Videos are going onto www.youtube.com/@ComplexityT... which will later also feature student videos. We did Williams (/Cook-Mertz/Shalunov) TIME(t) in SPACE(~√t) today.

youtube.com

Complexity Theory At Carnegie Mellon

05717

Reposted by Juspreet Singh Sandhu

TCS+ @tcsplus.bsky.social · 10/09/2026

Behold! An announcement. Our first TCS+ talk of the season will be by a survey talk on average-case lower bounds by the excellent Tselil Schramm (Stanford), next week, 🗓️ September 16 at 1pm ET! Mark your calendar, get ready, and sharpen your best note-taking pencils!

095

Reposted by Juspreet Singh Sandhu

The Fields Institute @fieldsinstitute.bsky.social · 10/09/2026

🌌 What is math’s mysterious Langlands program really about? Hidden correspondences hint at a deeper structure to the mathematical universe. Read more as Quanta Magazine columnist Natalie Wolchover unpacks one of those connections and asks mathematicians what they might mean. ⬇️

quantamagazine.org

What Is Math’s Mysterious Langlands Program Really About? | Quanta Magazine

Hidden correspondences hint at a deeper structure to the mathematical universe. Our columnist unpacks one of those connections and asks mathematicians what they might mean.

022

Reposted by Juspreet Singh Sandhu

Simons Foundation @simonsfoundation.org · 10/09/2026

How does string theory inspire new mathematics? In this Q&A, Junior Fellow Elise LePage of the Simons Society of Fellows talks about creating beautiful math to tackle open problems in theoretical physics. Read more: bit.ly/4ysxWcf #math #science #physics

bit.ly

Elise LePage: How String Theory Leads to Beautiful Mathematics

Junior Fellow Elise LePage of the Simons Society of Fellows turns intuition from theoretical physics into new mathematics.

031

Juspreet Singh Sandhu @freegaussians.bsky.social · 10/09/2026

@ccanonne.github.io - Tagging you, because what @angelamczhou.bsky.social writes here is super important, and maybe you can spread it around too. If we don't get objective, 3rd-party audits to measure the technology capability *and* get ethical/legal compliance checks the situation will devolve.

020

Juspreet Singh Sandhu @freegaussians.bsky.social · 10/09/2026

I was initially skeptical, but given the way things have evolved over the last years (and the concerns about plagiarism and academic misconduct from the labs in particular) — I believe this is the right thing to do.

021

Juspreet Singh Sandhu @freegaussians.bsky.social · 09/09/2026

@ccanonne.github.io and I doing some back of the envelope calculations to understand the cost-and-benefit analysis of oAI’s recent (dubious, and possibly academic-integrity violating) NS resolution.

010

Juspreet Singh Sandhu @freegaussians.bsky.social · 08/09/2026

@jnshi.bsky.social and I have worked hard to 1) build a deep program, and 2) slowly train & introduce new students to become strong and work on it (new paper w/ a student will be out soon). As Angela says: the long-term consequences of eroding certain ethics/ways of inquiry/training will be severe.

120

Juspreet Singh Sandhu @freegaussians.bsky.social · 08/09/2026

1) Kudos to Tristan-Apolge for extending (with help from LLMs) the technical machinery developed by Cordoba-Zoroa’s breakthrough work to getting to a path that can go all the way to NS: mastodon.social/@tristanbuck... 2) Many thanks to Terry for a good overview: mathstodon.xyz/@tao/1172335... 1/2

mastodon.social

tristanbuckmaster (@tristanbuckmaster@mastodon.social)

Today, Levent Alpöge and I have made public three results: finite-time blowup with smooth forcing for incompressible porous media, for Boussinesq, and for 3d incompressible Euler. https://cims.nyu.ed...

1121

Juspreet Singh Sandhu @freegaussians.bsky.social · 08/09/2026

I suppose the rate at which Lopatto is making progress, little will *not* be understood about Parisi measures for p-spin mixtures. He has already shown the validity of the full dAT line, fRSB for SK at any beta > 1, and here he constructs a mixture that exhibits k-RSB for any integer k.

001

Reposted by Juspreet Singh Sandhu

Marius Lemm @mariuslemm.bsky.social · 15/10/2025

Are the dynamics of disordered quantum many-body systems easier to simulate? --> Yes, because entanglement spreads slowly for a non-perturbatively long time (a.k.a. "prethermal MBL"). This is true in any dimension. arxiv.org/pdf/2510.11831 Thanks to my collaborators Ben McDonough and Andy Lucas!

arxiv.org

152

Reposted by Juspreet Singh Sandhu

arXiv math.PR Probability @mathpr-bot.bsky.social · 04/09/2026

Emmanuel Kammerer, Konstantinos Kavvadias, Jason Miller, Yi Tian: The conformally invariant metric on CLE$_4$ III: uniqueness arxiv.org/abs/2609.04140 arxiv.org/pdf/2609.04140 arxiv.org/html/2609.04140

011

Reposted by Juspreet Singh Sandhu

arXiv math.PR Probability @mathpr-bot.bsky.social · 04/09/2026

Emmanuel Kammerer, Konstantinos Kavvadias, Jason Miller, Yi Tian: The conformally invariant metric on CLE$_4$ II: existence of geodesics arxiv.org/abs/2609.04139 arxiv.org/pdf/2609.04139 arxiv.org/html/2609.04139

011

Reposted by Juspreet Singh Sandhu

arXiv math.PR Probability @mathpr-bot.bsky.social · 04/09/2026

Emmanuel Kammerer, Konstantinos Kavvadias, Jason Miller, Yi Tian: The conformally invariant metric on CLE$_4$ I: subsequential limits of the non-simple CLE graph metric arxiv.org/abs/2609.04138 arxiv.org/pdf/2609.04138 arxiv.org/html/2609.04138

011