Anthropic merges Cowork into the Claude chat interface
Anthropic has folded Cowork into chat, letting the model decide whether to answer directly or run a long task in the background, which the author says reshapes pricing and compute routing.
Anthropic just erased the line between asking AI a question and handing AI a job. Until now that line was product architecture. Chat in one app, agents in another, and you picked the door based on whether you needed a 10-second answer or 40 minutes of background work. That split existed because models couldn't route their own work. Someone had to judge how much autonomy a request deserved before typing it. The product org made you that someone. Merging Cowork into chat moves the judgment inside the model. You type one sentence, close your laptop, and Claude decides whether it owes you an answer or an afternoon of work. Watch what this does to pricing. A single text box that spans 5-second replies and hour-long jobs means the real question shifts from "which model" to "how much compute does this deserve." Whoever owns that routing decision owns the margin on every task. Every lab that shipped a separate agent product will end up here. Separate apps were a confession that the model couldn't tell a question from a job. That confession just expired.
This post quotes another post: x.com/1943306828697550848/status/210025849074053…