Reproduce “Avatar Intelligence: Amplifying Human Agency Through Persistent Personal Agents” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zheng Wu, Xinbei Ma, Yao Yao, Zhaoyuan Huang, Haowen Hu, Yijie Lu, Taolei Wang, Yifei Yang and 3 more
The authors propose Avatar Intelligence, a persistent-agent paradigm for delegating decisions while preserving user control. The system would carry a user model across tasks, act within delegated authority, and seek clarification when needed. They propose AOI, combining net user value with an oversight penalty. The paper offers a roadmap, not an implementation or evidence that oversight is reduced.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.