Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: Lowfat – pluggable CLI filter that saved 91.8% of my LLM tokens

github.com

Post reply on HN
Print view View original

91–97 of 97 posts

Re: Show HN: Lowfat – pluggable CLI filter that saved 91.8% of my LLM tokens

#91

Post by naorsabag »

zdkaster wrote: post #18
threecheese wrote: post #10

The docs are missing any examples of what this does, instead showing _how_ it works - and only for the codebase itself, rather than the behavior of the app. What would be useful: - examples of text that can be filtered, and why that would be valuable - a data flow diagram of runtime behavior, showing how filtering removes unnecessary context

Thanks for your feedback. Will put this in place. Meanwhile, please checkout architecture doc and plugin. The plugin doc could a little bit giving insight of what it does.

Agreed, try to use OpenHop to create the data flow diagram

Re: Show HN: Lowfat – pluggable CLI filter that saved 91.8% of my LLM tokens

#92

Post by urax »

tuo-lei wrote: post #28

the bigger problem is agents defaulting to the broadest command possible. kubectl get -o yaml when a jsonpath query would give 1/50th the tokens. filtering after the fact works, but you're still paying for the round trip. better to teach the agent to ask narrow questions in the first place.

[flagged]

Re: Show HN: Lowfat – pluggable CLI filter that saved 91.8% of my LLM tokens

#95

Post by Bnjoroge »

hboon wrote: post #86
Bnjoroge wrote: post #83

Earlier quoted context omitted.

definitely doesnt is a strong word. it technically is possible, but you might get banned

That was true. But actually, I think that's changed a few weeks ago since they introduced a API credit amount equivalent to your (eg. $100, $200) that will be used for such cases. So they don't ban you, they just bill you that allocated credit and then actual API cost.

Yes. That’s possible in addition to using your actual subscription. I’ve been using it via cliproxy for all harnesses and even my own code review agent hooked up to github apps. Not banned yet but I also dont do crazy stuff with openclaw or hermes

Re: Show HN: Lowfat – pluggable CLI filter that saved 91.8% of my LLM tokens

#96

Post by GodelNumbering »

smallerize wrote: post #75
onlyrealcuzzo wrote: post #58

Earlier quoted context omitted.

> The defaults are fine, but you can do better. Why can I do better than Pi? I don't want to build my own harness and deal with the bugs... I want to build my project... My understanding is that Codex / Claude / Gemini subscriptions don't work with custom harnesses. It's pretty hard to beat 5x more usage if you have the $200/mo subscription by using the API instead.

If you're looking for an efficiency-focused harness, I had a pretty good time using the Dirac agent. The line-based anchors were slightly buggy though (this was a couple months ago) and would sometimes add the same line of code multiple times or leave an anchor in the output.

Creator of Dirac here. Came across this too late. The behavior you mentioned happens more commonly in smaller models, rarely in larger/frontier models. The underlying code is clean but the smaller models often make boundary errors (off by one type). Which models did you observe this with?

Re: Show HN: Lowfat – pluggable CLI filter that saved 91.8% of my LLM tokens

#97

Post by smallerize »

GodelNumbering wrote: post #96
smallerize wrote: post #75

Earlier quoted context omitted.

If you're looking for an efficiency-focused harness, I had a pretty good time using the Dirac agent. The line-based anchors were slightly buggy though (this was a couple months ago) and would sometimes add the same line of code multiple times or leave an anchor in the output.

Creator of Dirac here. Came across this too late. The behavior you mentioned happens more commonly in smaller models, rarely in larger/frontier models. The underlying code is clean but the smaller models often make boundary errors (off by one type). Which models did you observe this with?

That tracks, this was a quantized version of GLM 4.7 Flash. I was surprised because I though this specific thing was not supposed to happen lol.
Post reply on HN