Show HN: Continual Learning with .md (github.com)

34 points by wenhan_zhou ↗ HN
I have a proposal that addresses long-term memory problems for LLMs when new data arrives continuously (cheaply!). The program involves no code, but two Markdown files.

For retrieval, there is a semantic filesystem that makes it easy for LLMs to search using shell commands.

It is currently a scrappy v1, but it works better than anything I have tried.

Curious for any feedback!

18 comments

[ 164 ms ] story [ 509 ms ] thread
I really like the simplicity of this! What's retrieval performance and speed like?
Minimalism is my design philosophy :-)

Good question. Since it is just an LLM reading files, it depends entirely on how fast it can call tools, so it depends on the token/s of the model.

Haven't done a formal benchmark, but from the vibes, it feels like a few seconds for GPT-5.4-high per query.

There is an implicit "caching" mechanism, so the more you use it, the smoother it will feel.

I've seen a lot of such systems come and go. One of my friends is working on probably the best (VC-funded) memory system right now.

The problem always is that when there are too many memories, the context gets overloaded and the AI starts ignoring the system prompt.

Definitely not a solved problem, and there need to be benchmarks to evaluate these solutions. Benchmarks themselves can be easily gamed and not universally applicable.

Context bloat is real, but the architecture has the potential to solve it.

You need clever naming for the filesystem and exploration policy in AGENTS.md. (not trivial!)

The benchmark is definitely the core bottleneck. I don't know any good benchmark for this, probably an open research question in itself.

[flagged]
The editability is surely an underrated advantage, both for the program itself and the memories it generated.

I think in terms of noise, it is less problematic here because not everything is being retrieved. The agent can selectively explore subsets of the tree (plus you can edit the exploration policy by yourself).

Since there is no context bloat, it is quite forgivable to just write things down.

I love how you approached this with markdown !

I guess the markdown approach really has a advantage over others.

PS : Something I built on markdown : https://voiden.md/

Seems interesting. Ill give it a try on my agent, memory is definitely an ongoing issue. How long have you been running this in a continuous state? Also have you tried other LLM's and seen a difference on how well they can use it?
Your example is with Codex - OpenAI could implement this easily on their end right? Every prompt of yours was an API call and they have a log, they can easily re-create a quick history of what you did/asked for before?
Have you noticed an relationship between recall and the number of files/memories?
Using the same approach for dev documentation storage: https://ctxlayer.dev/

Has been working for me for a couple of months already. So far human curation of context is the way to go.