Claude just unified their async workflow system. Cowork and chat modes are now merged into a single interface.
The core upgrade: context persistence across sessions. You can drop a task or question, close everything, and Claude continues processing in the background. When it needs clarification, it prompts you asynchronously rather than blocking.
You maintain veto rights on all outputs before they finalize. Think of it as supervised background execution rather than fully autonomous agents.
Rollout: Pro and Max tiers over the next few weeks. Team tier probably follows later.
This is basically Anthropic's answer to ChatGPT's persistent threads, but with built-in async task handling. Could be useful for long-running research tasks or multi-step workflows where you don't want to babysit the terminal.
The core upgrade: context persistence across sessions. You can drop a task or question, close everything, and Claude continues processing in the background. When it needs clarification, it prompts you asynchronously rather than blocking.
You maintain veto rights on all outputs before they finalize. Think of it as supervised background execution rather than fully autonomous agents.
Rollout: Pro and Max tiers over the next few weeks. Team tier probably follows later.
This is basically Anthropic's answer to ChatGPT's persistent threads, but with built-in async task handling. Could be useful for long-running research tasks or multi-step workflows where you don't want to babysit the terminal.