4 ms·

That'd be really easy to spot and also fix, most likely. Any open issue you could point us to, must surely been reported already?

nolok 5mo ago
> That'd be really easy to spot and also fix, most likely Ah, reminds me of good old "There are only 2 hard problems in computer science: cache invalidation, naming things, and off-by-1 errors."
- criemen 5mo ago
  > Ah, reminds me of good old "There are only 2 hard problems in computer science: cache invalidation, naming things, and off-by-1 errors." You quip, but LLM KV caching (from the harness side) is quite easy: You get a cache hit on stable prompt prefixes, period. That means you want to keep the prefix stable, and only append at the end of the conversation. Made up example: Don't put the git branch name into the system prompt part (that comes first), as whenever the branch name changes, that'd trigger a cache invalidation of the entire prompt. Getting this right requires some care to not by accident modify the prefix, basically, and some design on communicating the things that can change (user configuration, working dir, git information, ...).
  - xcjsam 5mo ago
    [dead]
  - franknord23 5mo ago
    That sounds like the experience of writing Containerfiles; since steps are cached you want to pull the thing you are iterating on as far down as possible.
    - gopher_space 5mo ago
      All of this work has been done before in different contexts. Memory management with bigger blocks and weaker definitions that change whenever some grad student gets a bright idea.
      - vidarh 5mo ago
        100%. Since you mention memory management: Generational GC is pretty much the same idea: Keep the stuff that's least likely to change an important property (liveness) together. Conceptually the underlying general idea is to sort things based on stability if you can avoid recomputing properties of the stable part.
    - verdverm 5mo ago
      It's even closer to prefix matching on super long strings by chunk
krzyk 5mo ago
Opencode (and other coding agents) have hundreds of open issues reported. It is quite discouraging when they are not being closed/fixed.
- jsjsjsuduiwkw 5mo ago
  [dead]
- verdverm 5mo ago
  These projects have also been the recipients of PR spam, lots of duplicates and unconfirmed in there for less technical people and clawd operators
3uler 5mo ago
https://github.com/anomalyco/opencode/pull/14743 https://github.com/anomalyco/opencode/pull/14743
- polski-g 5mo ago
  Yeah that's frustrating. Everytime they pop up on my Twitter timeline I shame them into addressing their issue backlog. They're aware of the issues length and they're "looking into a solution".

nolok 5mo ago

> That'd be really easy to spot and also fix, most likely Ah, reminds me of good old "There are only 2 hard problems in computer science: cache invalidation, naming things, and off-by-1 errors."

criemen 5mo ago
> Ah, reminds me of good old "There are only 2 hard problems in computer science: cache invalidation, naming things, and off-by-1 errors." You quip, but LLM KV caching (from the harness side) is quite easy: You get a cache hit on stable prompt prefixes, period. That means you want to keep the prefix stable, and only append at the end of the conversation. Made up example: Don't put the git branch name into the system prompt part (that comes first), as whenever the branch name changes, that'd trigger a cache invalidation of the entire prompt. Getting this right requires some care to not by accident modify the prefix, basically, and some design on communicating the things that can change (user configuration, working dir, git information, ...).
- xcjsam 5mo ago
  [dead]
- franknord23 5mo ago
  That sounds like the experience of writing Containerfiles; since steps are cached you want to pull the thing you are iterating on as far down as possible.
  - gopher_space 5mo ago
    All of this work has been done before in different contexts. Memory management with bigger blocks and weaker definitions that change whenever some grad student gets a bright idea.
    - vidarh 5mo ago
      100%. Since you mention memory management: Generational GC is pretty much the same idea: Keep the stuff that's least likely to change an important property (liveness) together. Conceptually the underlying general idea is to sort things based on stability if you can avoid recomputing properties of the stable part.
  - verdverm 5mo ago
    It's even closer to prefix matching on super long strings by chunk

krzyk 5mo ago

Opencode (and other coding agents) have hundreds of open issues reported. It is quite discouraging when they are not being closed/fixed.

jsjsjsuduiwkw 5mo ago
[dead]
verdverm 5mo ago
These projects have also been the recipients of PR spam, lots of duplicates and unconfirmed in there for less technical people and clawd operators

3uler 5mo ago

https://github.com/anomalyco/opencode/pull/14743 https://github.com/anomalyco/opencode/pull/14743

polski-g 5mo ago
Yeah that's frustrating. Everytime they pop up on my Twitter timeline I shame them into addressing their issue backlog. They're aware of the issues length and they're "looking into a solution".