Anthropic has updated Claude's memory so that what you tell the chatbot in an ordinary conversation carries over into Claude Cowork, and what happens in Cowork carries back into chat.
According to Adgully, the change is aimed at reducing the need for users to repeatedly re-explain themselves — restating a project's details, a writing style, or a set of preferences every time they switch surfaces. Crypto Briefing describes the same move as Anthropic unifying Claude's memory across chat and Cowork.
The wrinkle is the default. As Yahoo Tech's report puts it, Claude and Cowork "will share memories about you now — unless you opt out," framing the merge as one that raises privacy questions about what an AI remembers and whether the convenience is worth the tradeoff. Coverage from SiliconANGLE points to customization controls sitting alongside the improved memory, suggesting users are not stuck with the default.
Why it's a meaningful shift: memory has quietly become the competitive battleground for AI assistants. A model that forgets you is a tool you operate; a model that remembers you across contexts starts behaving more like a colleague who already knows the account, the deadline, and how you like things phrased. That is genuinely useful, and it is also the exact moment the privacy calculus changes — an offhand remark in a casual chat can resurface inside a work session you did not expect it to touch.
The sources here do not specify a rollout timeline, which regions or plans are covered, or precisely how the opt-out is surfaced, so those details remain open.
It matters because the default setting on a feature like this quietly decides how much of your life an AI assistant is allowed to carry from one room into the next.