NewBizz AI Ready: your AI server, delivered installed and configured.
See the serviceResources / News
Cursor's agent can now click and type in desktop applications and test the code it writes.
The agent in the code editor Cursor has gained a "Computer Use" feature. According to AlphaSignal, the agent can now click and type in desktop applications.
According to the report, the agent can test the code it writes using tools such as the iOS Simulator: it does not just write the code, it tries the result on screen.
AI agents are moving from producing text to doing work on screens and in applications. The more an agent can touch, the more it matters what access it is given. In enterprise use this raises three questions: what can the agent reach, at which step is human approval required, and is everything it does logged?