Tag
Chinese researchers achieved a breakthrough in autonomous AI agent engineering with a code-as-harness paradigm that replaces text prompts with executable verification substrates, enabling deterministic multi-agent execution through six internal processes.
This survey paper from Meta, Stanford, and Illinois argues that AI agents perform better when code is used as their primary working layer, treating code as the environment for reasoning, action, and modeling. The authors introduce the concept of an 'agent harness' encompassing tools, memory, sandboxes, and feedback loops.