Self-Learning Systems Advance AI Reasoning Following NousCoder-14B Release

Researchers are developing self-play systems like Absolute Zero Reasoner to help AI models surpass human training through automated problem generation and refinement.The breakthrough comes as Nous Research launched its open-source NousCoder-14B model, which achieved high accuracy on LiveCodeBench using Nvidia B200 processors.Collaborative projects involving Tsinghua University, Meta, and Salesforce aim to create superintelligence by allowing models to question and learn independently.