RESOLVING THE LEDGER
This is one outlet's own report from Engadget — the article as it was filed. Other outlets are covering the same event; open the full story to compare every source side by side.
See the full story · 4 sources
Plus, Anthropic is giving users more control over what memories Claude saves.
Anthropic Since last summer, Claude has had the ability to remember past conversations , but often those "memories" have been fragmented. Sometimes by design , and at other times due to limitations in Claude's architecture. That's partly changing today, with Anthropic releasing a major update to its memory feature set.
The company's models can now access and carry over memories across both Claude the chatbot and Claude Cowork. Cowork, if you need a reminder, is the computer use agent Anthropic released at the start of the year . You can use it to automate rote tasks you repeat often through the day. "When Cowork runs a task in the cloud, what Claude remembers from your chats is there, and vice versa," says Anthropic. In practice, this means you'll need to repeat yourself less often when using Cowork since Claude will have access to context from your chats. Today's update is well-timed since it builds on a recent upgrade Anthropic released for its Claude for Chrome browser extension where it made Cowork available through the plugin. Since that change, conversations you have in the Chrome side panel now become part of your usage history.
Anthropic is also giving users more control over Claude's memory. Over the last few weeks, users have been able to see the chatbot's memories organized by topic, and edit the specifics of each topic from Claude's settings menu. To that the company is adding a new "include sensitive topics in memory" option. "By default, Claude does not store topics related to personal or sensitive subject matter, like your health, race, ethnicity, religious beliefs, politics, gender identity, and other similar areas," explains Anthropic.
If you flip on the sensitive topics toggle, Claude will begin adding those topics to its memory. Each time it does so, you will see a notification. Anthropic notes Claude won't retroactively add past conversations to its memory, and that you can turn off the feature at any time, as well as edit and delete what it saves. As for why it's making this change now, the company reasons "what some consider sensitive, others may consider useful for Claude to remember." Anthropic notes Claude will not save certain information, including your social security number, criminal history, immigration status, even if you have the sensitive topics toggle enabled.
Memory is available (and enabled by default) for all Free, Pro and Max accounts. On mobile, update to the latest version of the Claude Android or iOS app to see the new updates.
AIPROPX is an independent multi-source news index — we track, compare, and connect coverage from across the web into one place you won't find anywhere else.