frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Show HN: Is grep enough? A transparent benchmark for agentic code navigation

https://entelligentsia.github.io/is-grep-enough/
2•bonigv•1h ago
Felt LSP Servers were too complex. Bash tools alone too brutish. Wanted to see what if it is a tree-sitter as a firstclass tool. Ran a bench over 10 large codebases [bitcoin, django, rails, redis,...] at 5 levels of exploration complexity each. That 150 context isolated runs over the last few days. Sharing the results with full tarnsparency. All scripts, docker image scripts, all transcrpts. There is a TL;DR; but I hope you don't leave it at that. Has been quite a bit of work. Repo links are on the site.

Comments

6thbit•1h ago
This is nicely put together, it does make sense that lsps help more as complexity grows because makes navigation across symbols easier.

I hope someone with a large budget can reproduce these with latest Opus/gpt.

My gut feeling is that higher reasoning models tend to use grep more effectively. But intuitively lsp should still win there.

bonigv•57m ago
You are absolutely right about what we feel intuitively - LSPs should beat the shit out of the competition. But surprisingly it did not. Across 10 different LSP servers, across 5 different levels of prompt complexity it did not. Mind you, I painstakingly warmed up the LSP servers that needed it warmed. Some liked it cold and it fared equally non impressively. The pattern I saw was, LLMs (sonnet w.6 with cc) was very clever to use whatever it had to get to a verifiable answer. It could do it just with bash for sure. But as the prompt complexity grew the cost also rose.

Treesitter is sitting in a sweet spot here. a vrainy LLM can find the shortest path with high quality with treesitter and a few bash calls.

Api.weather.gov's robots.txt disallows all bots

https://api.weather.gov/robots.txt
1•mikeocool•1m ago•0 comments

Cursor for iOS

https://cursor.com/blog/ios-mobile-app
1•benjlang•1m ago•0 comments

Art Benefits Transaction (ABT): A Proposal for Economic Stimulus

https://www.reddit.com/r/PublicPolicy/s/fCXl3HefIX
1•taivare•5m ago•1 comments

Has Perfume Become Samey?

https://buchananb.substack.com/p/has-perfume-become-samey
1•0x54MUR41•6m ago•0 comments

Big Data File Formats

https://luminousmen.substack.com/p/big-data-file-formats
1•Tomte•6m ago•0 comments

Sunwæe – your life's AI OS

1•dvdxnss•8m ago•0 comments

Language Design Impacts Security

https://www.adacore.com/blog/how-language-design-impacts-security
1•ajdude•8m ago•0 comments

Why Token Optimization Is a Gift to the Hyperscalers

https://www.uncoveralpha.com/p/why-token-optimization-is-a-gift
1•gmays•9m ago•0 comments

Help! My passive fund is aggressively US tech focused

https://monevator.com/help-my-passive-fund-is-aggressively-us-tech-focused/
1•rocco8620•10m ago•0 comments

The Humanoid That Pays to Stand Still – Robotics

https://atomsfrontier.substack.com/p/the-humanoid-that-pays-to-stand-still
1•jpatel3•10m ago•0 comments

Open USD

https://joinopenstandard.com/blog/introducing-open-usd
2•coloneltcb•11m ago•0 comments

Show HN: TraceAIO – open-source LLM visibility tracker

https://traceaio.org
2•owenthejumper•11m ago•0 comments

Solved.Earth

https://Solved.Earth
2•benheaton•11m ago•0 comments

Counterfeit Verifiability in Autonomous Agent Payments

https://zenodo.org/records/21042364
1•adamzwasserman•11m ago•0 comments

Speculative Supply Chains: How Rational Incentives Manufacture Madness of Crowds

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7022818&__cf_chl_f_tk=v0gWvo.uXho1ZsUoo8e07qq...
1•petethomas•11m ago•0 comments

The AI Productivity Trap

https://i0exception.substack.com/p/doing-more-shipping-less
1•i0exception•12m ago•0 comments

Does Social Media Use Matter for Students' Well-Being?

https://link.springer.com/article/10.1007/s10902-026-01070-y#Sec29
1•mpweiher•12m ago•0 comments

Show HN: Statuslin.es – a community library of custom Claude Code status lines

https://statuslin.es
1•nastynate•14m ago•0 comments

Using Playwright to test my static sites

https://alexwlchan.net/2026/playwright/
2•surprisetalk•14m ago•0 comments

AI and Us: It's Complicated

https://syntheticauth.ai/posts/ai-the-falsity-of-comparison
1•zerolayers•14m ago•0 comments

Workers' share of income explains why many Americans are down on the economy

https://www.cbsnews.com/news/labor-share-income-lowest-since-world-war-ii/
1•ripe•17m ago•0 comments

Reasoning About Async Rust with State Machines

https://aibodh.com/posts/async-rust-chapter-2-what-async-fn-compiles-into/
1•febin•17m ago•0 comments

HTTP Status Codes Explained (100–599)

https://urlwatch.io/blog/http-status-codes.php
1•mssblogs•17m ago•2 comments

Mojo Quest: A browser-based game for learning Mojo syntax

https://quest.mojolang.org/
1•mdunnoconnor•19m ago•0 comments

Too many tables are bad for you

https://www.cybertec-postgresql.com/en/too-many-tables-are-bad/
1•0x54MUR41•19m ago•0 comments

Fata Morgana (Mirage)

https://en.wikipedia.org/wiki/Fata_Morgana_(mirage)
1•keiferski•19m ago•0 comments

Rendering ray tracing in a database (ClickHouse)

https://github.com/ClickHouse/RayTracer
2•sdairs•20m ago•0 comments

Out of the loop

https://saturnino.substack.com/p/out-of-the-loop
1•pramodbiligiri•20m ago•0 comments

The Grammar of Data: Define Once, Run Anywhere with Cross-Engine Expressions

https://xorq.dev/blog/grammar-for-data-engineering/
1•zazuke•20m ago•0 comments

Show HN: Debategle – ranked 1v1 debates judged by an LLM

https://debategle.com/
1•sawsymikey•20m ago•0 comments