Reddit devs compare agent memory approaches: summaries, MEMORY.md, wikis and vector DBs all fail differently
ombre-ai · reddit · 2026-11-06
A Reddit discussion on cross-session agent memory compares four common approaches—running summaries, MEMORY.md-style files, wiki/notes folders, and vector DB/memory services—and how each fails: summaries compress away reasons, a single file buries key lines as it grows, wikis go stale without maintenance, and retrieval resurfaces outdated decisions. The author asks which approach holds up for coding, research, writing, and ops work, what people have dropped, and how stale decisions stop getting used.
More from coding & agent
- ~780 AI slop accounts identified on X; users deploy agents to auto-block them daily — TheZachMueller · 2026-11-06
- Formal proofs of tens of millions of lines of code are now feasible, says Anthropic engineer — morqon · 2026-11-06
- Terminal-Bench's two broken tasks traced to Debian 11 archive repo migration — PawelHuryn · 2026-11-06
- Stable AI proposes DPDM: frozen LLM representations plus shared predictor boost agent decisions — stable-ai · 2026-11-06
- One prompt ran Codex for 17+ hours with 109 sub-agents and 2B tokens — andimarafioti · 2026-11-06
- MCP vs API vs CLI: build both on one operations layer, prioritize MCP — hottown · 2026-11-06