Show HN: Watch a neural net learn to play Snake
Summary
A web-based tool that visualizes a neural network (using PPO) learning to play Snake in real-time, with configurable parameters and 3D rendering.
View Cached Full Text
Cached at: 05/15/26, 06:33 PM
Similar Articles
@neural_avb: This is what you can achieve with 5-6 hours of Self-Play RL training by the way Actors view the projectiles with lidar …
A thread sharing a video of self-play RL training with lidar and PPO in Unity, followed by a lecture on building AlphaGo from scratch.
I built a site that lets you watch, wager, and prompt inject agents playing games
A developer built a site where users can watch AI agents play games, wager fake coins, and use winnings to prompt inject agents. The author shares observations about model performance, noting that smaller models struggle while Qwen3 235B excels.
Deep Neural Network that turns any Image into a Playable Game ! All on consumer GPUs and Not Datacenters
The author presents a small transformer-based neural network trained from scratch to turn any image into a playable game, running in real-time on consumer GPUs like an RTX 5090. The model uses autoregressive decoding with KV caching but currently has issues with motion and context.
Learning to play Minecraft with Video PreTraining
OpenAI introduced Video PreTraining (VPT), a semi-supervised method that trains neural networks to play Minecraft by learning from 70,000 hours of unlabeled human gameplay video combined with a small labeled dataset. The model learns complex sequential tasks using the native human interface (keyboard and mouse) and demonstrates capabilities like crafting diamond tools and pillar jumping, representing progress toward general computer-using agents.
I built a chess coach that explains moves like a grandmaster instead of showing engine lines — powered by LLM
A chess coaching tool that uses an LLM to explain moves in natural language like a grandmaster, replacing raw engine evaluations with contextual coaching narratives. It analyzes user games from Chess.com and Lichess with local Stockfish, detects recurring mistakes, and offers personalized chat and spaced repetition for blunders.