A RL-trained microduck learns to slide like a penguin
kevin_zakka · x · 2026-09-02
Henry J. Brew shared a reinforcement learning demo where a small character (microduck) learns to slide on its belly like a penguin. The motion looks natural and fluid — a fun showcase of RL-driven physics animation.
More from Fun
- Google Slides 'Beautify' feature sparks complaint over unnecessary design intervention — mmitchell_ai · 2026-09-02
- Grok misjudges user location as Singapore due to WiFi connection — suchenzang · 2026-09-02
- User Showcases Professional Short Animation Made with Minimax-h3 — solomars3 · 2026-09-02
- Open source plea: Closed video models hinder adoption — VoidAsuka · 2026-09-02
- Satire on 'frontier model' definition: Only SF EA region qualifies? — beffjezos · 2026-09-02
- Dev humor: 'User-Agent: python-requests' making 10k POST requests — djcows · 2026-09-02