Anthropic upgrades Claude Code/Cowork to control Macs in background

This move signals a shift from AI as a chat interface to AI as an ambient operator of personal hardware, which could reshape productivity workflows for knowledge workers and developers alike. By enabling background control, Anthropic is betting on trust in autonomous agents for routine, multi-step tasks—raising practical questions about error handling, oversight, and system reliability when the user is not actively supervising. The restriction to paid tiers suggests this is a premium capability, likely to pressure competitors to match or differentiate. What to watch is whether background control expands beyond Macs and whether usage data reveals meaningful gains in task completion without a rise in unintended actions.