Researchers build self-improving game environments to train AI agents

Original: awesome work by the PI team 👏

Source: x.com ↗