Hacker News Viewer

Show HN: Hippo, biologically inspired memory for AI agents

by kitfunso on 4/6/2026, 9:49:34 PM

https://github.com/kitfunso/hippo-memory

Comments

by: nberkman

Cool project. I like the neuroscience analogy with decay and consolidation.<p>I&#x27;ve been working on a related problem from the other direction: Claude Code and Codex already persist full session transcripts, but there&#x27;s no good way to search across them. So I built ccrider (<a href="https:&#x2F;&#x2F;github.com&#x2F;neilberkman&#x2F;ccrider" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;neilberkman&#x2F;ccrider</a>). It indexes existing sessions into SQLite FTS5 and exposes an MCP server so agents can query their own conversation history without a separate memory layer. Basically treating it as a retrieval problem rather than a storage problem.

4/6/2026, 10:37:06 PM


by: kami23

Cool to see others on this thread.<p>Here&#x27;s a post I wrote about how we can start to potentially mimic mechanisms<p><a href="https:&#x2F;&#x2F;n0tls.com&#x2F;2026-03-14-musings.html" rel="nofollow">https:&#x2F;&#x2F;n0tls.com&#x2F;2026-03-14-musings.html</a><p>Would love to compare notes, I&#x27;m also looking at linguistic phenomena through an LLM lens<p><a href="https:&#x2F;&#x2F;n0tls.com&#x2F;2026-03-19-more-musings.html" rel="nofollow">https:&#x2F;&#x2F;n0tls.com&#x2F;2026-03-19-more-musings.html</a><p>Hoping to wrap up some of the kaggle eval work and move back to researching more neuropsych.

4/6/2026, 11:16:13 PM


by: swyx

hmm the repo doesnt mention this at all but this name and problem domain brings up HippoRAG <a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2405.14831" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2405.14831</a> &lt;- any relation? seems odd to miss out this exactly similarly named paper with related techniques.

4/6/2026, 11:24:17 PM


by: the_arun

Aren&#x27;t tools like claude already store context by project in file system? Also any reason use &quot;capture&quot; instead of &quot;export&quot; (an obvious opposite of import)?

4/6/2026, 10:40:49 PM


by: matt765

cool project mate, gj

4/7/2026, 12:32:33 AM


by: gfody

yegge has a cool solution for this in gastown: the current agent is able to hold a seance with the previous one

4/6/2026, 11:07:09 PM


by: cyanydeez

no open code plugin? This seems like something that should just run in the background. It&#x27;s well documented that it should just be a skill agents can use when they get into various fruitless states.<p>The &quot;biological&quot; memory strength shouldn&#x27;t just be a time thing, and even then, the time of the AI agent should only be conformed to the AI&#x27;s lifetime and not the actual clock. Look up <a href="https:&#x2F;&#x2F;stackoverflow.com&#x2F;questions&#x2F;3523442&#x2F;difference-between-clock-realtime-and-clock-monotonic" rel="nofollow">https:&#x2F;&#x2F;stackoverflow.com&#x2F;questions&#x2F;3523442&#x2F;difference-betwe...</a> monotonic clock. If you want a decay, it shouldn&#x27;t be related to an actual clock, but it&#x27;s work time.<p>But memory is more about triggers than it is about anything else. So you should absolutely have memory triggers based on location. Something like a path hash. So whever an agent is working and remembering things it should be tightly compacted to that location; only where a &quot;compaction&quot; happens should these memories become more and more generalized to locations.<p>The types of memory that often are more prominent are like this, whether it&#x27;s sports or GUIs, physical location triggers much more intrinsics than conscious memory. Focus on how to trigger recall based on project paths, filenames in the path, file path names, etc.

4/6/2026, 10:13:16 PM


by: esafak

How does it select what to forget? Let&#x27;s say I land a PR that introduces a sharp change, migrating from one thing to another. An exponential decay won&#x27;t catch this. Biological learning makes sense when things we observe similar things repeatedly in order to learn patterns. I am skeptical that it applies to learning the commits of one code base.

4/6/2026, 11:27:03 PM


by: bambushu

[dead]

4/6/2026, 11:47:40 PM