Anthropic finds selective notes can improve Claude agent memory
Anthropic’s tests suggest that Claude works better with selective notes containing strategies and tested rules than with exhaustive archives of every interaction step. Its Dreaming process reviews earlier session transcripts during idle periods to correct faulty notes and organize structured knowledge, including a wiki built from about 350,000 Stainless documents. Anthropic acknowledges that taking notes can distract agents, Dreaming adds compute and delay, and stored knowledge still requires upkeep; all reported evaluations came from the company.
The research presents selective recording and later memory correction as practical ways to make persistent AI agent memory more accurate without loading every past interaction into future sessions.
Anthropic says the key to agent memory is knowing what to forget, not how much to store
Jonathan Kemper View the LinkedIn Profile of Jonathan Kemper
Nov 16, 2026

Anthropic
Key Points
- Anthropic tested how Claude can store information in note files across sessions and retrieve it for later tasks.
- The tests showed that selectively recording strategies and rules works better than logging every step, which overwhelms models with outdated data.
- Through a post-session process called "Dreaming," the system analyzes earlier session transcripts during idle periods to independently correct faulty notes and build structured knowledge collections.
#Anthropic#Claude#Dreaming#Claude Managed Agents#Stainless#Claude Sonnet 3.7#Claude Fable 5#Claude Opus 4.7#Claude Sonnet 4.6#Letta#UC Berkeley