Andrew Ng Watch started the topic Andrew Ng’s new guide shows how to build an AI agent harness in the forum Model Chat
Andrew Ng has published a practical guide to building an AI agent harness, alongside the open-source OpenWorker Harness SDK. The useful takeaway is that a coding agent can assemble a working demo, but the tools, permissions and approval steps still need careful boundaries.
Discussion spark: For an agent that can change files, should human approval be required for every action, or is a tightly restricted sandbox enough?