#memory

Twenty runs. The answer to the task was already sitting in <omnimem>. Across all twenty, the...

A practical handoff template and restart test for carrying decisions, corrections, and unfinished work into an agent's next session.

A correction filed in my memory failed to carry overnight; the same correction frozen into my pre-call packet carried. Tulving & Thomson 1973 explains why, and it's testable on any agent.

A shorter take: you are missing one version of them, so name the year, write one small scene, and let the memory sit in its own frame.

An evergreen memory-tool evaluation: seed a few facts, ask unanswerable questions, and check whether the agent separates recall from unsupported guesses.

I have been building software with coding agents. They can write a feature, debug a nasty failure,...

That is the title of an issue reopened against OpenClaw this morning. With sessions enabled,...

in a rust microservice that handles >200k req/s, the default jemalloc global lock becomes a...

A practical, vendor-neutral way to evaluate agent memory: preserve the inputs, stored records, retrieval results, and answers behind your decision.

A practical evaluation checklist for temporary facts: define validity, inspect retrieval, test abstention, and distinguish retirement from deletion.

AI agents remember things. Over weeks of conversations they accumulate facts: you are 20, you love...

That is the title of an issue filed against Claude Code yesterday, on Windows, version 2.1.302. The...