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

Grep beats LSP? Why coding agents ignore your fancier tools

agentconnect.md

Post reply on HN
Print view View original

71–79 of 79 posts

Re: Grep beats LSP? Why coding agents ignore your fancier tools

#71

Post by genxy »

I see no one has referenced the written by ai paper itself https://arxiv.org/abs/2608.13568

https://github.com/agentconnect-md/lsp-vs-grep-token-study

The author uses their own harness, I'd want to see if what they are seeing is related to the harness, they should at the minimum also use CC which has direct LSP support.

Their metric is token economy, many of us don't pay per token and capability in terms of difficulty and quality are a more important metrics for the kinds of work I do. I'd want those tested as well.

Re: Grep beats LSP? Why coding agents ignore your fancier tools

#72

Post by genxy »

poytr1 wrote: post #41
tonyarkles wrote: post #8

There’s been an interesting co-evolution that I’ve been experiencing with Claude Code. I’ll ask it to do a task, I’ll watch what it’s doing (often lots of find and grep and ripgrep) and then after the task is complete I’ll ask it if there are any tools that would’ve made the job easier. This has led to tools like fzf and others (notmuch for indexing email, for example). I’ve then taken those tools and figured out how…

I had a similar intuition that we could definitely optimize token efficiency and performance for non-general tasks. For example, I was handling a lot of infra operations, many of which follow fixed workflows like locating a specific log in certain time range, identify the issue, and apply the appropriate mitigation steps. Since these workflows are entirely command-line based, I added a hook to Claude Code that monito…

You should disclose that you are the co-author of this paper.

Re: Grep beats LSP? Why coding agents ignore your fancier tools

#73

Post by Volundr »

polotics wrote: post #52
N_Lens wrote: post #24

Earlier quoted context omitted.

It’s me and my plucky gang of agents!

Nah as long as you don't give them first-names, I think you should be ok. also: Pretty pleaser please people don't get your agents registered as direct-reports in the org-chart with HR!

But how else am I supposed to collect their salary?

Re: Grep beats LSP? Why coding agents ignore your fancier tools

#74

Post by czottmann »

zfy0701 wrote: post #67

anyone had experience with https://github.com/colbymchenry/codegraph ?

It's good, and my agent is using it just fine (explicitly told it to prefer it over just grepping randomly). I use it in Swift, Go, and Node projects. If you use worktrees, make sure to run `codegraph init` post-setup to avoid confusing the robot :)

Re: Grep beats LSP? Why coding agents ignore your fancier tools

#76

Post by Bjartr »

bandrami wrote: post #60
norman784 wrote: post #58

Earlier quoted context omitted.

Not only Lisp, but Javascript and Ruby (AKA dynamic languages) devs also rely heavily in REPLs, for me at least the reason is that you don't know what value a variable holts, so you need to inspect them somehow, with typed languages you have less reasons to reach such tool.

iPython was famous for a while but I think its popularity has waned, but yes as you point out really any dynamic language can do a repl; it's just that it's mostly the lisp/scheme and forth families that have leaned so heavily into the idea.

Not only dynamic languages! Java has `jshell` for instance

Re: Grep beats LSP? Why coding agents ignore your fancier tools

#77

Post by lirolero »

NewsaHackO wrote: post #17
tonyarkles wrote: post #8

There’s been an interesting co-evolution that I’ve been experiencing with Claude Code. I’ll ask it to do a task, I’ll watch what it’s doing (often lots of find and grep and ripgrep) and then after the task is complete I’ll ask it if there are any tools that would’ve made the job easier. This has led to tools like fzf and others (notmuch for indexing email, for example). I’ve then taken those tools and figured out how…

When you use “we” hopefully you are referring to someone on your team. Just be careful, the first step down the rabbit hole of AI psychosis is humanization of LLM.

Is that how you got into this state?

Re: Grep beats LSP? Why coding agents ignore your fancier tools

#78

Post by fooofw »

> On a noisy TypeScript repository, it improved F1 by 0.246 and used 12% fewer tokens. The useful predictor was lexical noise, not whether the language had strong static types.

What is "lexical noise", which appears to be a key term in the argument, and why is it not defined? Does it mean that an identifier could show up in, say, a docstring in a non-structured way where a semantic tool would not identify it?

Makes sense to me that LLMs, which are language based, would match better with a tool that searches at a language lever rather than at a deeper structured level.

Post reply on HN