Andreas Longva

I've moved back to Norway with my family after resigning from my position at RWTH Aachen University. I've been on medical leave since Jan 2024, as a result of what likely is a variant of Long COVID. I hope to somehow eventually continue my research and wrap up the final project for my PhD. 1/4

New video is up! Why is the simplex method so fast? At STOC'26 we presented the best theory yet

Our Q3 2026 technical report is out! During the past three months we introduced significant performance and feature improvements to our Rapier and Nexus physics engines: soft-bodies, MPM, C and python bindings, and more! dimforge.com/blog/2026/10... #rapier3d #rapier2d #nexus2d #nexus3d

I'm one of the candidates for this year's Eurographics Executive Committee election! If you are an EG member, please make sure to vote :) My platform: - No AI in papers or reviews - Keeping what we like about Eurographics: the topics, the review quality, the scale. We won't become another ICLR

Spending my Saturday satisfying my ICLR reciprocal reviewing duty, which means sifting through hundreds of pages of completely meaningless AI slop. What is my moral / professional obligation as a reviewer when faced with a meaningless AI-written submission?

New blog post on speeding up the separating axis test (SAT) using inscribed spheres: box2d.org/posts/2026/0...

hi friends I'm very excited to announce Halfspace, my experimental cross-platform IDE for solid modeling with distance fields! This has been in progress since midway through last year, and it's finally "done enough" to make it more public.

As graphics researchers or geometry teachers, we speak about 3D using screenshots, and that's just sad... What if you could make SLides on top of PolyscOPE? You can using SLOPE! Slope allows you to synchronize a 3D viewer, animated meshes, shaders, live demos, equations (& more) with your slides!

Lame old me talked a bit (25min) about the role of visualization in science --- forgive my poor time management: youtube.com/watch?v=7CfDEnHxPWA At BCON I was personally interested to witness the constant practical constraints surrounding performance, and editability/directability.

Can we please start holding AI companies criminally liable for their AI agents hacking various companies — and now even the healthcare system in Australia? Why on earth are they not held responsible?

A huge milestone: Fearless SIMD 1.0 is published: linebender.org/blog/fearles.... This was a real collaboration, thanks especially to Sergey Davidoff for pushing this over the line, Laurenz Stampfl, Andrew Jakubowicz, valadaptive, Benjamin Saunders, and Daniel McNab.

AI leaders have you worrying about P(doom), but it's really P(dystopia) you should be concerned about

It’s been 8 years since we published our report on how Google tricks people into extensive location tracking, and simultaneously filed complaints against Google. Today the decision from the Irish Data Protection Commission arrived: a €400 million euros fine. www.forbrukerradet.no/siste-nytt/d...

Cursed Statue

Katastrofale tall for norske skoleelever i Pisa-undersøkelsen. Er det mulig å forklare dette med noe annet enn satsingen på iPad og "digitalisering" i stedet for skolebøker? La oss i hvert fall ikke gjøre den samme feilen ved å introdusere KI i skolen

#geogram just reached 2500 stars on github, And that's not all, version 1.10.1 was just released: github.com/BrunoLevy/ge... Geogram is in C++, but it also supports Python, Node.js, WebAssembly, TypeScript and Lua through Rosetta by Frantz Maerten, more on this here: github.com/rosetta-bind...

Announcing the Palomar registry of Lean formalized mathematics: palomar-registry.org . See also my blog announcement at terrytao.wordpress.com/2026/08/18/p... and the Lean Zulip channel at leanprover.zulipchat.com#narrow/chann....

Excited to announce: In 2027 I'll join the UW Allen School in Seattle as tenure-track faculty in Computer Science! My group will advance core geometry processing + 3D AI for visual computing, science, & engineering. I'll be recruiting PhD students in the next cycle!

Charles Babbage intentionally designed the computer as an automation of mental labor. Today, computer graphics automates visual art production via privately governed generative models. How did we get here? Critical Labor History of Generative Models, SIGGRAPH Talks '26 🧵

Google AI still doing about 50/50 on basic facts in my experience. Kinda crazy that they just don't care. I see way too many people using Google AI overviews for fact checking. Please don't! Go to trusted sources instead. The skills to verify information have never been more important.

So.... A good concavity measure is just an unsolved problem in geometry?

The talk I presented at #SIGGRAPH2026 is now available on YouTube, for anyone who couldn’t attend or would like to revisit it :) youtu.be/OFJYBUArVPc?...

Have you dug deep into an evolving area of visual computing? Consider submitting to the 2027 EG STARs track! STARs are survey reports providing insight into the current state of the art and future directions. They're impactful and highly-cited! eg2027.isti.cnr.it/call-for-sta...

At #SIGGRAPH2026, I presented "points as tori" (PAT) for fast pointwise evaluation of signed distance for point clouds. Some code (Python w/ C++ extensions) and demo up: github.com/nzfeng/point... Contributions from people/LLMs welcome. Explainer thread below for the project: 🧵

Today at #siggraph2026, Disney Animation has four more talks! Here's a mini-thread about our four talks today, including three from the Hyperion rendering team about Disney Animation rendering tech: (1/10) disneyanimation.com/events/?draw...

boy I’m sure glad we solved AI alignment before deploying modes that could autonomously jailbreak themselves and hack cloud provider infrastructure stuff could have gotten really weird

Next week at #SIGGRAPH2026, we'll present "Fast Sparse Matrix Permutation for Mesh-Based Direct Solvers," with Behrooz Zarebavani, Ana Dodik, Changcheng Yuan, Serban Porumbescu, John Owens, Maryam Mehri Dehnavi, and @justinmsolomon.bsky.social . Talk: Wed., July 22, 3:55 PM, Room 403-B (1/6)

After some encouragement from various people I decided to make a GitHub repo for my ongoing Newton's Cradle benchmark of Google's Veo video model. All videos (spanning about a year) are available here for all your cradling needs: github.com/dilevin/newt...

When we presented GravoMG, a fast multigrid solver for surfaces, the first question we got was: can you make this work for tetrahedral meshes? Tomorrow, Marcel Padilla will present the answer at SMI in Istanbul. Sneak preview: yes, it's really fast. Code and paper: marcelpadilla.com/Projects/Gra...

Congratulations to Bernhard for his SIGGRAPH 2026 Paper 👍 PDF ge.in.tum.de/download/ST-... , Video youtu.be/-1Txagqj4N0 Our key idea: treat particles as samples in four-dimensional space-time. ST-FLIP acts as a temporal anti-aliasing mechanism for FLIP-style solvers.

I’m happy to announce the release of a new open source 3D physics engine called Box3D. I’ve been working on this project for a few years now, but it represents over 20 years of experience writing physics engines for games. Read more here: box2d.org/posts/2026/0...

Sheldon Andrews (profs.etsmtl.ca/sandrews/) is running for the ACM Executive Committee this year. I've known Sheldon for a long time and I think he will do a great job ! I voted for him, you should too www.acm.org/elections/si...

I recently completed my PhD thesis, where I developed algorithms for making geometric inferences from imperfect data (i.e. incomplete and inaccurate observations/representations of shapes, like broken meshes or point clouds). A thread 🧵:

Giving a talk in the @Stanford SCIEN seminar this Wednesday (1/3) at 4:30pm: scien.stanford.edu/index.php/ev... The topic is “normal coordinates”: a shape representation little-used outside of mathematics—but which turns out to have nice applications in geometry processing & learning.

New paper by the incredible Donglai Xiang (xiangdonglai.github.io) at CVPR 2026 using RKPM inside a data free training framework to compute reduced elastic models. Trains faster, works better than competing Neural Network approaches. Check it out: research.nvidia.com/labs/sil/pro...

Got my daughter to eat a whole plate of veggies because she wants to become big enough to reach the next handle at the playground. She wants to swing like Spider-Man 😁

When dealing with noisy and/or partial inputs to describe a shape, the standard approach in geometry processing is to reconstruct a clean implicit surface first, and to do geometry on it. But why this one in particular? What about all the other surfaces that would have been coherent with the input?

We have a fun new paper to appear at SIGGRAPH 2026, led by Columbia PhD student Xiana Carrera, that extends dual contouring to work nicely with [pure] signed distance samples. It avoids classic DC's additional requirement of providing accurate Hermite / […] [Original post on mastodon.acm.org]

Plouf ! Our @eurographicsc.bsky.social course notes on Optimal Transport for Fluids, New and Old, with @pltc.bsky.social are out, here: diglib.eg.org/items/68a801...

Having a wonderful time at the Eurographics Conference in Aachen! I gave the longest talk of my life (3h) on the biggest screen of my life and I received the Eurographics PhD award for my thesis 🎉 A huge thanks to @eurographicsc.bsky.social and all my collaborators and supporters!