What happened
Amazon Web Services (AWS) introduced four Nova 2 AI models, enhancing price-performance across reasoning, multimodal, conversational, code generation, and agentic tasks. Concurrently, Nova Forge was launched, enabling customers to build optimised Nova variants by integrating proprietary data at pre-trained, mid-trained, and post-trained stages through an 'open training' approach. Nova Act, powered by a custom Nova 2 Lite model, achieved 90% reliability for browser-based UI automation workflows.
Why it matters
The Nova Forge 'open training' approach introduces a tightened dependency on customer data integration across the entire model training lifecycle, increasing the burden on data engineering and governance teams to manage data provenance and quality for custom AI models. This raises due diligence requirements for legal and compliance teams concerning intellectual property and data security when infusing proprietary datasets into foundational models. The high reliability of Nova Act for UI automation shifts the operational burden to validating and integrating these automated processes.
Subscribe for Weekly Updates
Stay ahead with our weekly AI and tech briefings, delivered every Tuesday.
