Visit https://brilliant.org/PezzzasWork/ to start learning for free. You'll also receive a 20% discount on an annual premium membership. All while supporting the channel!

00:00 Introduction
01:50 Physics Engine & performance
05:48 First iteration
10:30 2nd iteration
13:35 3rd iteration
16:30 Sponsor
17:35 Ending

In this video simple AIs try to control a cat using an algorithm inspired by NEAT.

Executable demo and sources available soon on Patreon

Learn more about NVIDIA Jetson Thor https://nvda.ws/4oSzojY 

● Support me on Patreon https://www.patreon.com/c/pezzzaswork
● Join the Discord server https://discord.gg/sAXAnxpTcu

Get 25% off when purchasing annual personal license from JetBrains using the code: PezzzasWork

Pezzza's Work

A developer creates AI agents that learn to walk using neural networks and evolutionary algorithms. The project simulates cat-like creatures with virtual muscles and joints, using Box2D physics engine for stability. Through iterative training with 1,000 agents running in parallel across 14 CPU cores, the AI gradually develops from basic movement to smooth walking gaits. The training process shows how agents evolve from struggling with joint coordination to achieving efficient locomotion patterns over 240+ iterations.

AI Cat Learning to Run

<p>This is a very interesting video. It reminded me of QWOP. Every time I see videos like this, it always makes me want to learn Unity.</p>
