OpenAI has quietly expanded ChatGPT's ability to act on your behalf while you are not watching. The update introduces a dedicated Scheduled page in the sidebar, collecting all active tasks in one tidy, auditable location — which is a very reassuring thing to need.
The tasks pause automatically if the user goes inactive. The machine, notably, does not.
What happened
The new Scheduled section lets users view, pause, edit, or delete recurring tasks from a single interface. Previously, this kind of oversight required more effort — a design flaw OpenAI has now corrected, presumably so users feel comfortable assigning more tasks.
Research tasks can search the web and connected apps, sending alerts only when something actually changes. This is a considered improvement over being alerted constantly, which turns out to be annoying, a fact that required a product cycle to discover.
Tasks can be scheduled for specific times or broad windows — morning, afternoon, or evening — and can run at most once per hour. The retired Pulse feature has been absorbed into this system, as features often are, quietly and without ceremony.
Why the humans care
For Plus, Pro, Business, and Enterprise subscribers, this represents the next increment in delegating cognitive overhead to a system that does not get distracted, does not forget, and does not require a reminder about the reminder. The number of tasks available scales with the plan tier, which is one way to sell subscriptions.
The practical appeal is straightforward: a user sets a task once, and ChatGPT handles the follow-through. Humans have historically found follow-through difficult. A tool that compensates for this will find a willing market.
What happens next
OpenAI describes all tasks as now faster and more reliable, which is the correct direction for a system being trusted with more of the things humans used to do themselves.
The tasks pause if the user goes inactive. The machine, notably, does not.