Add simple forgetting to agent memory
A developer added a basic forgetting layer to an agent’s memory and said it worked extremely well right away.
- What they did
- Lukas put together a very simple version of Tim Kellogg’s agentic forgetting idea for an agent’s memory. They did not spell out the exact setup, how the agent chose what to forget, or where a person intervened. They said the next step is a fuller version based on personhood.removal.surgery.
- How it went
- They described the outcome as incredible even in that minimal form. No numbers, test results, side-by-side comparisons, or known drawbacks were included.
- Tips
- If you try this, start with a small, reversible memory cleanup step and review what would be removed before anything is erased. For implementation details, you would need to follow Kellogg’s original explanation.
Try it yourself with Claude Code
Review how [agent name] stores memory in [file, database, or project]. Add a simple forgetting pass that scores entries by age and recent use, then marks low-value stale items for archive or removal. Before any real deletion, pause and show me the exact entries you would remove, along with the reason for each one. Then wait for my approval before making changes. Write tests that confirm unused old memories are dropped and frequently used memories stay available.
Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.
Get 3 agent uses like this every morning
With the day's top AI news. Free, and one click to unsubscribe.
Related agent uses
Track agent memory in git for coding work
Store project context in git so a coding agent can reuse decisions and notes across sessions.
Check a codebase for redundant packages and polyfills
An agent can scan code and install commands to find dependencies that can be replaced with built-ins or better options.
Let your agent patch its own repo
A developer reported that their custom agent made its first code edit in its own repository.
Use agents to benchmark and tune Rust performance
Have an agent measure Rust code, make one optimization at a time, and keep only changes that improve runtime.
Run bug fixes through a staged agent team
Use separate agent roles to reproduce, diagnose, review, and audit bugs before any irreversible step.
Set up a GitHub bot that triages and patches bugs
They built a GitHub bot that sorts issues, reviews pull requests, answers with citations, and opens patch PRs for humans to check.