SachitRafa

38 ms·

1.
by SachitRafa 3mo ago
I agree to your point, in our current system we are not treating each memory as same. Based on the importance, category and recall frequency we decide a decay rate which would not be same for all the memories. A architectural decision will
2.
by SachitRafa 3mo ago
The final artifact will look like a flat scored memory graph, in which each node will be a memory having threshold between 0-1 which will be derived through various parameters such as recall, importance etc. Edges will connect semantically
3.
Show HN: YourMemory, agentic memory is a pruning problem, not a hoarding problem(yourmemoryai.vercel.app)
19 points by SachitRafa 3mo ago|3 comments
4.
by SachitRafa 4mo ago
Release link : https://github.com/sachitrafa/YourMemory/releases/tag/v1.4.2...
5.
by SachitRafa 4mo ago
Repo Link : https://github.com/sachitrafa/YourMemory
6.
Show HN: YourMemory, persistent memory layer with temporal reasoning for agents
9 points by SachitRafa 4mo ago|4 comments
7.
by SachitRafa 4mo ago
So based on the importance that it gives the reports individual will decide that which report will survive in the system. You can manually also identify and set the importance or can let llm be the deciding factor here to assign weightage.
8.
Show HN: Fixing AI memory blind spot on connected facts with benchmark(yourmemoryai.xyz)
7 points by SachitRafa 4mo ago|2 comments
9.
by SachitRafa 4mo ago
How would you know if the agent is not querying stale data here ?
10.
by SachitRafa 4mo ago
Would be definitely trying this, how do you check the rollback of API calls which would not be an part of a transaction?
11.
by SachitRafa 5mo ago
Sure will have a look !
12.
by SachitRafa 5mo ago
Hi All, Based on the feedback, have made few new enhancement. 1. Activity aware decay, now instead of a wall clock we run decay based on active days in which the user has been using the system. 2. Spatial/ Working directory memory, now
13.
by SachitRafa 5mo ago
Wall clock decay punishing vacation is a issue. The current state of a clock decay is for simplicity but I do see how a session based decay might be helpful. For failures and strategies it still might work as env drift on calendar anyhow (n
14.
by SachitRafa 5mo ago
Hi Builder here, biological is here doing the rhetorical work. The actual mechanism here is exponential decay with category specific half lives, recall based reinforcement and pruning context which falls below threshold. The main difference
15.
by SachitRafa 5mo ago
Comingling unrelated projects is a fair argument for this. But the goal is here not to cater to specific project patterns but general standards you want to use across project. Using MD files for this is fine till a point. If you keep on add
16.
by SachitRafa 5mo ago
No it's not based on softmax output. It's single pass for now !
17.
by SachitRafa 5mo ago
Also if any of you are curious on how to set it up. You only need to execute these 2 commands pip install yourmemory yourmemory-setup
18.
by SachitRafa 5mo ago
Hi all builder here, quick update. Benchmarked it against LongMemEval dataset and got the result 84.8% for recall top 5 and 86.8% for nDCG any@5. All the methodology and results are mentioned in the repo !
19.
Show HN: AI memory with biological decay (52% recall)(github.com)
98 points by SachitRafa 5mo ago|53 comments
20.
by SachitRafa 5mo ago
Yes this is the way forward in CICD as we move more towards agentic ai development more and more, great project !
21.
by SachitRafa 5mo ago
All memories should not decay at the same rate. these weights should be configurable !
22.
by SachitRafa 5mo ago
What is this solving actually and how is this different from a private insta account ?
23.
by SachitRafa 5mo ago
This is hilarious and oddly soothing. Tolkien is the ultimate stress test for long term context if the agent describes landscapes forever, how are you preventing it from looping or losing the "world state"? I’ve been obsessed with
24.
by SachitRafa 6mo ago
The name could have been better
25.
by SachitRafa 6mo ago
The CD-ROM streaming approach is the real insight here, keeping only activations and KV cache in RAM and streaming weights one matrix at a time sidesteps the 32MB constraint entirely. It's essentially the same trick modern edge inferen
26.
by SachitRafa 6mo ago
The cross-realm argument for packages like is-string is the one I find hardest to dismiss, but even there the math doesn't add up. The number of projects actually passing values across realms is tiny, and those projects should be the o
27.
by SachitRafa 6mo ago
The payments background makes total sense as an origin for this I've seen teams reach for Ray or Prefect and spend weeks configuring things they didn't need, when the actual problem was just 'how do I get work off this one ma