Pinned Tweet
Behind every "100% written in Python" is a "56% written in C++"
(This is not an attack on the Genesis authors. Genesis is really cool and @taichi_lang is first in their list of acknowledgments. People just focus too much on the 100% Python aspect of it.)
WTF?! New open-source physics AI engine absolutely insane! 🤯 Genesis is a new physics engine that combines ultra-fast simulation with generative capabilities to create dynamic 4D worlds for robotics and physics.
TL;DR:
🚀 430,000x faster than real-time physics simulation, processes 43M FPS on a single RTX 4090
🐍 Built in pure Python, 10-80x faster than existing GPU solutions like Isaac Gym
🌐 Cross-platform support: Linux, MacOS, Windows, with CPU, NVIDIA, AMD, and Apple Metal backends
🧪 Unified framework combining multiple physics solvers: Rigid body, MPM, SPH, FEM, PBD, Stable Fluid
🤖 Extensive robot support: arms, legged robots, drones, soft robots; supports MJCF, URDF, obj, glb files
🎨 Built-in photorealistic ray-tracing rendering
⚡ Takes only 26 seconds to train real-world transferrable robot locomotion policies
💻 Simple installation via pip: pip install genesis-world
🤝 Physics engine and simulation platform are fully open-sourced
🔜 ”.generate” method/generative framework coming soon.
finally tried OMP, was getting frustrated my hermes agent and @thegeneralist01 had been shilling hard. i get it know. it's so good.
to me, this is the best part of the update.
I have avoided writing parallel code because you just couldn't abort it without loosing your kernel
I disagree with the concept of stated vs revealed preferences, because my revealed preferences are that I really fucking like re-compiling cuda
efebic retweeted
It's like a bunch of paleonthologists who cannot get rid of their opinions on dinosaurs, formed by thinking about bones, after they got literally resurrected in Jurassic Park.
I think it's sad
Everyone: wow! It just learned a new class without any training samples, it’s like magic.
The softmax function:
Geoffrey Hinton did an MNIST subliminal learning experiment in the 2015 paper 'Distilling the Knowledge in a Neural Network'!
They trained a neural network on a dataset that completly lacked the number 3 and it exactly what the number 3 looked like anyway.
What's even crazier is that you can train a neural network entirely on nosie data with ghost logits that dont correspond to any real numbers and have it learn to classify all the numbers!
It's cool when a phrase's literal and figurative translation line up. make's you think that some metaphors are just natural. here being irritated is being likened to fur being stroked in the wrong way in both English and Japanese.
