Sometimes it’s hard to tell if I’m fighting the model architecture or the undocumented shortcuts in the PTX ISA. At least document the truncation 🙃

When you forgot the flyback diode across the solenoid and the inductive kickback fried your MOSFET during the dry run

📢📢📢 𝐒𝐩𝐡𝐞𝐫𝐢𝐜𝐚𝐥 𝐕𝐨𝐫𝐨𝐧𝐨𝐢: directional appearance as a differentiable partition of the sphere – Project: https://t.co/mUPYvL887s – Paper: https://t.co/v1SZrVfVoA – Colab: https://t.co/WLJDMpv7Yu Project led by 𝐅𝐫𝐚𝐧𝐜𝐞𝐬𝐜𝐨 𝐝𝐢 𝐒𝐚𝐫𝐢𝐨 (graduating!)

EgoX: Generate immersive first-person video from any third-person clip A novel framework from KAIST AI & Seoul National University that leverages video diffusion models to transform a single exocentric video into a realistic egocentric view. See it in action!

@markgurman The average IQ of both companies has increased.

gonna miss my asian gang

Sharing an interesting recent conversation on AI's impact on the economy. AI has been compared to various historical precedents: electricity, industrial revolution, etc., I think the strongest analogy is that of AI as a new computing paradigm (Software 2.0) because both are

Added highly-scalable loop closure and bundle adjustment to my ultra-efficient VIO. See me running around my apartment for a few loops and return to starting point. Uses model on NPU instead of the classic bag-of-words; which is not very scalable. This is now VIO + Loop

BIG NEWS: Introducing SDF-Splats! 🚨 I just solved one of the biggest problems in 3D Gaussian Splatting: GAPS & HOLES! 🕳️❌ My hybrid approach combines: ✨ Explicit Gaussian Splats (for detail) ✨ Implicit SDFs (for continuity) The result? Seamless 3D reconstruction with NO

A haiku about my life: I am so tired Where did all my money go My back is hurting

How do we reconstruct a 3D scene from photos with varying exposures? Standard methods often fail, leaving you with blown-out colors or disturbing shadows. We're excited to introduce Neural Exposure Fields (NExF), our new work accepted at #NeurIPS2025! 🧵

This is legitimately mind-blowing... How the FUCK does Sora 2 have such a perfect memory of this Cyberpunk side mission that it knows the map location, biome/terrain, vehicle design, voices, and even the name of the gang you're fighting for, all without being prompted for any of

fyi, if your circle is mostly celebrating murder, you are the bad guy

ViPE: Video Pose Engine for 3D Geometric Perception Contributions: • A robust and efficient framework, ViPE, for estimating camera parameters and dense depth from diverse, in-the-wild videos. • A system design that integrates the strengths of classical SLAM (efficiency,

There have been a lot of crazy many-camera rigs created for the purpose of capturing full spatial video.  I recall a conversation at Meta that was basically “we are going to lean in as hard as possible on classic geometric computer vision before looking at machine learning

A research paper details how "proxies" can become a core interaction concept for mixed reality. Details here: https://t.co/OtNJstaJCJ

We’re releasing the Epstein Files day one. There are no Epstein Files. Here’s video proof nothing happened. Ok we had to edit it. There are no Epstein Files. There are Epstein Files, but the Democrats made them. Why do you care so much about this Epstein guy? What are “Files”?

https://t.co/rL6OgsqA0G

The Great Voltage Wave PCB Art by Sinchronous Electrics

SF is the only place in the world where coding at a Bar is socially encouraged.