Anthropic just shipped memory across conversations. Here's how I'm using it for my studies and content workflow.
Anthropic shipped memory across conversations and it changed my daily Claude workflow more than any feature this year. The short version: Claude can now remember preferences, ongoing projects, and context you've shared, without you re-explaining yourself every chat.
I have separate projects for HANZHI.ID, my videos, and my coursework. Memory means Claude knows my tone, the people on my team, and the boring constraints I would otherwise paste in every prompt.
How I actually use it
First, I let Claude save the things I tell it more than twice. If I'm correcting tone, citing a teammate, or pinning a constraint, that goes into memory. The rest stays in the conversation.
Second, I treat memory like a small CRM. Names, context, what I'm working on, what I dropped. I read it weekly and prune anything that's no longer true.
What I'd avoid
Don't dump your whole life into memory on day one. Let it grow from real use. The signal-to-noise ratio matters more than coverage.
Written by Holly in Beijing. Notes get longer as I learn more. If you spot something off, find me in the footer.