DealForge autonomously sources, scores, and writes investment memos on venture deals. Stop manually hunting.

1,180+ deals tracked  ·  22 AI investment memos  ·  Updated daily

← Back to leaderboard

AI memory with biological decay (52% recall)

Show HN: AI memory with biological decay (52% recall)

40 AI Score
Show_hn other Added Apr 26, 2026

Details

Sector
other
Total Funding
$0
Last Round
$0

About

Most RAG setups fail because they treat memory like a static filing cabinet. When every transient bug fix or abandoned rule is stored forever, the context window eventually chokes on noise, spiking token costs and degrading the agent&#x27;s reasoning.<p>This implementation experiments with a biological approach by using the Ebbinghaus forgetting curve to manage context as a living substrate. Memories are assigned a &quot;strength&quot; score where each recall reinforces the data and flattens its decay curve (spaced repetition), while unused data eventually hits a threshold and is pruned.<p>To solve the &quot;logical neighbor&quot; problem where semantic search misses relevant but non-similar nodes, a graph layer is layered over the vector store. Benchmarked against the LoCoMo dataset, this reached 52% Recall@5, nearly double the accuracy of stateless vector stores, while cutting token waste by roughly 84%.<p>Built as a local first MCP server using DuckDB, the hypothesis is that for agents handling long-running projects, &quot;what to forget&quot; is just as critical as &quot;what to remember.&quot; I&#x27;d be interested to hear if others are exploring non-linear decay or similar biological constraints for context management.<p>GitHub: <a href="https:&#x2F;&#x2F;github.com&#x2F;sachitrafa&#x2F;cognitive-ai-memory" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;sachitrafa&#x2F;cognitive-ai-memory</a>

AI Score Reasoning

Heuristic score based on available signals. Funding: $0, Source: show_hn.

Source

Show_hn — View original →