Browser agents and computer use — companion videos

Anthropic

A two-minute Anthropic demo of Claude planning a small multi-app task — search the web, check Maps, drop a calendar invite — by driving the desktop directly. Useful contrast to the cloud-browser-only model in the Operator demo below and a good gut check on the article's point that computer-use agents work best on short, well-bounded chores rather than open-ended work.

What you should get from this: Compare desktop computer-use behavior with browser-only agents.

Watch or know first: Read the companion article's safety section first; the Operator demo below adds the browser-side picture.

OpenAI

The actual launch demo of Operator, where the team books restaurants, orders groceries, buys event tickets, and lets Operator stall on a redirect or hand control back when it hits a login. It is the clearest picture you will find of how a browser agent feels in practice — the screenshot-and-click loop, the confirmations before "stateful" actions, the prompt-injection guard rails — which is exactly the texture the article is trying to set expectations for.

What you should get from this: Recognize the browser-agent action loop, where it helps, and where human confirmation is still required.

Watch or know first: Basic understanding of agents and why live web tasks are harder than text generation.