A side project I've been building lately. I built a neural net with Tensorflow and trained it on the MNIST dataset of handwritten images. Then, I brought the model into Blender and used the values to turn it into a 3D scene. This is actually the model running!

Excited to share our work on Neural Assets: a new method for enabling 3D asset-level control in image diffusion models – scalable & without any 3D inductive biases. Neural Assets goes beyond text or pixel-based control & provides an interface inspired by 3D graphics tools. 🧵

I can't* fathom why the top picture, and not the bottom picture, is the standard diagram for an autoencoder. The whole idea of an autoencoder is that you complete a round trip and seek cycle consistency—why lay out the network linearly?

“Everyone knows” what an autoencoder is… but there's an important complementary picture missing from most introductory material. In short: we emphasize how autoencoders are implemented—but not always what they represent (and some of the implications of that representation).🧵

There is no word in English that means someone is 30% sure something will happen

I'm skeptical of the claims being made in this thread. Estimating a camera's lens blur field is a commendable technical achievement, but equating this to a fingerprint for reliable individual camera identification is irresponsible, and not supported by the research.

Generated 3D worlds aren't just for viewing, you can also play projectile jenga inside link below

📢 Lyra: Generative 3D Scene Reconstruction via Video Diffusion Model Self-Distillation Got only one or a few images and wondering if recovering the 3D environment is a reconstruction or generation problem? Why not do it with a generative reconstruction model! We show that a

Vibe coding is crazy man. Met a 12yr old making $600k a month with his vibe-coded SaaS he started 4 months ago. I asked how he built so fast. He said he just made a design-doc in GPT and fed it to Cursor with Sonnet-4 and it worked first try. His goal is to get to $2M a month

Title: Advice for a young investigator in the first and last days of the Anthropocene Abstract: Within just a few years, it is likely that we will create AI systems that outperform the best humans on all intellectual tasks. This will have implications for your research and

this is beyond mindblowing for me. somebody built a 5 million param language model inside minecraft, trained it, equipped it with basic conversational ability. probably the best thing i have seen entire month.

How Kernel Regression is related to Attention Mechanism - a summary in 10 slides. 0/1

6 years later, still the best introduction to pytorch internals.

Excited to release new repo: nanochat! (it's among the most unhinged I've written). Unlike my earlier similar repo nanoGPT which only covered pretraining, nanochat is a minimal, from scratch, full-stack training/inference pipeline of a simple ChatGPT clone in a single,

Lucas Atkins This code is extremely dangerous. Here, I improved it.

. Nano Banana Pro Create an image of 45°58′35″N 7°39′31″E at sunset

an IDE that shows you ads while your agent is spinning and the ads pay for the tokens

this is the most used slash command internally at cursor to remove ai slop install it to your project with this link: cursor.com/link/command?n….

Wondering how to attend an ML conference the right way? ahead of NeurIPS 2025 (30k attendees!) here are ten pro tips: 1. Your main goals: (i) meet people (ii) regain excitement about work (iii) learn things – in that order. 2. Make a list of papers you like

all roads lead to the same subspace "The Universal Weight Subspace Hypothesis" 500 ViTs, 500 Mistral LoRAs, 50 LLaMA models all collapse into shared 16-dimension subspaces regardless of data Not every problem in AI is a data one