How to Use Claude's Massive New Upgrades

What It Covers

Anthropic released five major Claude upgrades over roughly one month — Computer Use, Dispatch, Remote Control, Channels, and Scheduled Tasks — collectively shifting Claude from a chat tool into a persistent, always-on orchestration layer that executes multi-hour workloads autonomously across devices while users are away from their desks.

Key Questions Answered

• •Remote Control (Claude Code): Start a Claude Code terminal session on your desktop, then continue directing it from your phone via URL or QR code. The session runs entirely on your local machine — accessing your file system, MCP servers, and configs — while mobile acts purely as a window. Use slash-RC to switch mid-session without restarting.
• •Dispatch (Claude CoWork): A single persistent conversation thread on your phone that spawns and manages multiple simultaneous CoWork or Claude Code sessions on your desktop. One user directed roughly 25 minutes of async instructions across a full day, generating over three hours of parallel Claude execution — competitor analysis, Notion pages, infographic iterations — without sitting at a desk.
• •Computer Use: Claude can now control mouse, keyboard, and screen inside Claude CoWork, operating any desktop application without API integration. This makes it viable for legacy enterprise software with no native AI connectors. Combined with Dispatch, Claude can execute multi-app workflows autonomously while you are physically away from the machine.
• •Channels (Claude Code): Connect external event sources — Sentry alerts, CI failures, webhook payloads, Telegram or Discord messages — directly into a running Claude Code session. Unlike chat, channels push events in from outside, enabling Claude to react autonomously to monitoring triggers without user initiation. Designed for developers wanting a fully hackable, programmable event pipeline.
• •Scheduled Tasks: Claude CoWork and Claude Code both now support recurring scheduled tasks. Local schedules run while your machine is awake; cloud-based schedules run on Anthropic infrastructure with no local dependency. Practical use cases include sweeping open PRs, analyzing overnight CI failures, syncing documentation after merged PRs, and generating weekly spreadsheet updates automatically.