What is actually changing
Until now these were two different things, a chat window for questions and a separate Cowork space for longer agent tasks, and you had to know which one you wanted. That split is going away. You ask a quick question or you hand over a report, and Claude takes it from there, checking in with you when something isn't clear and leaving the final call to you. Cowork's agent abilities are moving into the main conversation, along with newer document and design features, so one Claude now covers both the fast answers and the slow, do-it-for-me work. This is less a new button than a change in what the product is for.
Where Anthropic is really headed
"Ask a quick question or hand over a report, and Claude takes it from there, even after you close your laptop." the company said.
The bigger play behind that line is consolidation. Anthropic has been running a set of separate Claude experiences, and it's now pulling them into one app it wants people to live in, the same superapp race Google and OpenAI are running too. Merging the casual chat and the autonomous agent is the clearest sign yet of that direction, because it treats "answer me" and "go do this" as the same product rather than two.
What has not been settled
The merge is going to Pro and Max subscribers first, over what the company only calls the next few weeks, so free users and the timeline for everyone else are still open questions. It also leaves the harder one unanswered, which is how comfortable people will be with an assistant that keeps acting on a task after they've walked away from the screen. Anthropic's answer, that Claude asks when it's unsure and you keep the final say, is the whole trust pitch, and it only gets tested once this is in real hands.