OpenClaw’s memory is unreliable, and you don’t know when it will break
blog.nishantsoni.com
OpenClaw’s memory is unreliable, and you don’t know when it will break
1–10 of 196 posts
Re: OpenClaw’s memory is unreliable, and you don’t know when it will break
#2I do feel like the memory the biggest hurdle I’ve been encountering and I’m curious what solutions people have been doing to make it work.
What seems to be somewhat working for me
1. Karpathy wiki approach
2. some prompting around telling the llm what to store and not.
But it still feels brittle. I don’t think it’s just a retrieval problem. In fact I feel like the retrieval is relatively easy.
It’s the write part, getting the agent to know what it should be memorizing, and how to store it.
Re: OpenClaw’s memory is unreliable, and you don’t know when it will break
#3I do feel like the memory the biggest hurdle I’ve been encountering and I’m curious what solutions people have been doing to make it work. What seems to be somewhat working for me 1. Karpathy wiki approach 2. some prompting around telling the llm what to store and not. But it still feels brittle. I don’t think it’s just a retrieval problem. In fact I feel like the retrieval is relatively easy. It’s the write part, ge…
As a side note, one thing that seemed to really help was having the agent recursively summarize all the files into a repo map with the same folder/file structure as the repo.
And forcing to always orient itself with that repo map first seemed to really help it from tunnel visioning.
Re: OpenClaw’s memory is unreliable, and you don’t know when it will break
#4"Who's in charge here?"
"The Claw."
Some of this stuff is starting to look like technologies that worked, looked promising, but were at best marginally useful, such as magnetohydrodynamic generators, tokamaks, E-beam lithography, and Ovonics.
Re: OpenClaw’s memory is unreliable, and you don’t know when it will break
#5Good to know that I'm not alone. I now use it for music recommendations (not so great) and keeping track of restaurants I want to try (really good at this but so are a lot of other apps).
Re: OpenClaw’s memory is unreliable, and you don’t know when it will break
#6Sounds like an armchair expert
Re: OpenClaw’s memory is unreliable, and you don’t know when it will break
#7Sounds like an armchair expert
Right back at ya ;)
The author makes some good conclusions; I’m as AI-pilled as the next hopefully-not-soon-to-be-ex-software-engineer, and I struggled to find use cases for my Claw that couldn’t be served with a cronjob and $harness.
If your findings contradict that, we are all ears - genuinely.
Re: OpenClaw’s memory is unreliable, and you don’t know when it will break
#8Could we stop with the clickbaiting headlines?
Re: OpenClaw’s memory is unreliable, and you don’t know when it will break
#9> 0 legitimate use cases
My teams currently using it for:
- SDR research and drafting
- Proposal generation
- Staging ops work
- Landing page generation
- Building the company processes into an internal CRM
- Daily reporting
- Time checks
- Yesterday I put together proposal from a previous proposal and meeting notes, (40k worth)
Re: OpenClaw’s memory is unreliable, and you don’t know when it will break
#10It would’ve happened eventually anyway, but OpenClaw is basically what kickstarted the beginning of the end of token subsidies. It’s a almost begging to be used wastefully. And agents would miss and lose nothing without it. It’s devoid of a reason to exist.