NVIDIA AI has unveiled Molt, a groundbreaking PyTorch-native framework designed to streamline agentic reinforcement learning (RL) research. Targeting the high iteration cost of algorithm modification, Molt offers a uniquely compact codebase, making it easier for human researchers and AI coding assistants to comprehend and evolve complex RL systems. This release promises to accelerate innovation in areas like multi-turn tool-use and LLM-as-judge applications.

Amazon SageMaker AI now offers serverless model customization for NVIDIA Nemotron 3 models, including Nemotron 3 Nano and Super. This powerful integration empowers enterprises to fine-tune high-performance, open-weight foundation models on domain-specific data, creating proprietary AI assets without managing complex infrastructure. The move significantly lowers the barrier to entry for specialized AI development, promising cost savings and enhanced data security.