Show HN: Lossless-memory – a personal AI memory that never summarizes
aru-labs
59 points
21 comments
September 21, 2026
Related Discussions
Found 5 related stories in 92.4ms across 7,302 title embeddings via pgvector HNSW
- Show HN: A public AI whose memory is shared across all users adjohu · 70 pts · August 16, 2026 · 68% similar
- Show HN: Loss. a tiny satire about AI progress pbobak · 30 pts · September 15, 2026 · 65% similar
- Show HN: Adaptive Recall, persistent memory for AI assistants over MCP abratabia · 20 pts · July 12, 2026 · 61% similar
- Show HN: I evaluated file, vector, graph and RL based memory frameworks pinglin · 13 pts · August 15, 2026 · 59% similar
- Show HN: MCP Memory – Fast Agent Memory Using Google's OKF and SQLite FTS5 pcbmaker20 · 58 pts · August 13, 2026 · 56% similar
Discussion Highlights (5 comments)
StilesCrisis
Is the date really useful? Usually I want the AI to remember rules, like "always do X before committing Y" and a timestamp doesn't change anything there.
alansaber
looks like a nifty information retrieval approach. cool. temporal/versioned chronology is absolutely useful.
rbansal2
Have you thought about using llm-wiki? It's pretty powerful and works well.
0xbadcafebee
It may seem useful at first, but eventually you'll hit a wall where this doesn't work, and you have to add another memory technique. Eventually you end up with a complex multi-layered system, because what people want by "memory" is actually 10 different things which all need their own solution.
theresLand
Seems like this would break the cache often. That would increase billing rates with certain providers and, for local models, take a while to generate responses, especially with long-running agentic sessions. Interesting idea though.