Tethered: Our Agents Are Us — Shu Fang, Two Sigma
Two Sigma, a quantitative hedge fund, gave all employees their own AI agent that runs under their individual identity instead of a general service account.
This solved major problems with permissions and licensing that arise when agents are separate from people. The main risk is that no one can distinguish what the agent did from what the person themselves did, so they use tracking headers to document every action. A larger security concern is web access — instead of giving the agent open internet, they use Google's enterprise search within their own network, which reduces the risk of sensitive data leaking.
Open on YouTube →
Vibekollen prepared this summary with AI from the original publication. The content belongs to AI Engineer.