📄 中文摘要
随着工作日的推进，知识工作者通常需要同时处理多个任务，如客户报告、预算表、幻灯片和电子邮件等。这些任务相互依赖，要求工作者同时关注。为了在这样的环境中真正发挥作用，AI 代理需要具备类似的操作能力。然而，目前最先进的模型仍然是以单一任务为基础进行评估，尚未实现多任务处理的能力。CORPGEN 的研究旨在开发更高效的 AI 代理，使其能够在复杂的工作场景中提供支持，提升工作效率。
CORPGEN 推进 AI 代理用于实际工作
出处: CORPGEN advances AI agents for real work
发布: 2026年2月26日
📄 English Summary
CORPGEN advances AI agents for real work
As the workday progresses, knowledge workers often find themselves managing multiple tasks simultaneously, such as client reports, budget spreadsheets, slide decks, and email backlogs. These tasks are interdependent and demand attention concurrently. For AI agents to be genuinely effective in such environments, they need to operate in a similar manner. However, today's best models are still evaluated on a single-task basis, lacking the capability for multi-task processing. CORPGEN's research aims to develop more efficient AI agents that can provide support in complex work scenarios, enhancing overall productivity.