Is Grep All You Need? How Agent Harnesses Reshape Agentic Search
Recent advances in Large Language Model (LLM) agents have enabled complex agentic workflows where models autonomously retrieve information, call tools, and reason over large corpora to complete tasks on behalf of users. Despite the growing adoption of retrieval-augmented generation (RAG) in agentic search systems, existing literature lacks a systematic comparison of how retrieval strategy choice interacts with agent architecture and tool-calling paradigm. Important practical dimensions, including how tool outputs are presented to the model and how performance changes when searches must cope with more irrelevant surrounding text, remain under-explored in agent loops. This paper reports an empirical study organized into two experiments. Experiment 1 compares grep and vector retrieval on a 116-question sample from LongMemEval, using a custom agent harness (Chronos) and provider-native CLI harnesses (Claude Code, Codex, and Gemini CLI), for both inline tool results and file-based tool results that the model reads separately. Experiment 2 compares grep-only and vector-only retrieval while progressively mixing in additional unrelated conversation history, so that each query is embedded in more distracting material alongside the passages that matter. Across Chronos and the provider CLIs, grep generally yields higher accuracy than vector retrieval in our comparisons in experiment 1; at the same time, overall scores still depend strongly on which harness and tool-calling style is used, even when the underlying conversation data are the same.
๐๐ ๐ฒ๐๐๐ซ๐ฌ ๐๐ ๐จ ๐ญ๐จ๐๐๐ฒ, ๐๐ง๐๐จ๐ ๐ฅ๐๐ฎ๐ง๐๐ก๐๐ ๐ฐ๐ข๐ญ๐ก ๐ ๐ฌ๐ข๐ฆ๐ฉ๐ฅ๐ ๐๐๐ญ: don't chase hypeโfollow the innovators & early adopters before they become the mainstream.
๐ ๐๐๐๐๐๐๐ฌ ๐ฅ๐๐ญ๐๐ซ: the tools have changed. The frameworks have changed. The architectural nightmares? Still here.
Our 20th-anniversary look at the technology adoption curve cuts through the noise, shows where the tech stack really sits today, and what's coming next:
https://bit.ly/4uVxdPg
๐ ๐๐๐ฉ๐ฉ๐ฒ ๐๐๐ญ๐ก ๐๐ง๐ง๐ข๐ฏ๐๐ซ๐ฌ๐๐ซ๐ฒ, ๐๐ง๐๐จ๐!
Suite de notre voyage ร Madrid avec Louis ๐ช๐ธ
Dรฉcouvrez le 2รจme volet de son retour sur la #MadVue2026, le rendez-vous des dรฉveloppeurs #VueJS. Au programme : Module Federation et Vue ร l'รจre de l'IA. ๐ค
๐ Lire l'article : https://buff.ly/lYZaqpm
Voici la seconde partie de mon retour sur la **MadVue 2026**, la confรฉrence Vue.js qui s'est tenue ร Madrid le 22 mai 2026. Si vous avez manquรฉ le dรฉbut, la premiรจre partie est juste ici : * <a href="https://www.wanadevdigital.fr/360-1-temps-reel-urls-typees-formulaires-et-toolchain-retour-de-la-madvue-2026/" target="_blank">Partie 1 - Temps rรฉel, URLs typรฉes, formulaires et toolchain</a> * <a href="https://www.wanadevdigital.fr/361-2-module-federation-et-vue-a-l-ere-de-l-ia-retour-de-la-madvue-2026/" target="_blank">Partie 2 - Module Federation et Vue ร l'รจre de l'IA</a> Aprรจs les fondations de l'application Vue moderne, on change d'รฉchelle et d'รฉpoque. Dans cet article, on va voir **comment partager des composants entre plusieurs applications Vue avec Module Federation**, puis trois talks consacrรฉs ร **Vue et l'intelligence artificielle** : faire **gรฉnรฉrer une interface en direct par un LLM**, **prรฉparer son projet Vue pour les agents**, et enfin **apprendre le frontend** ร l'รจre de l'IA. C'est parti !
No-code is expanding fast โ๏ธ
More teams can now automate workflows without technical expertise.
๐ More automation
โก Faster deployment
๐ More integrations
But increased automation also creates complexity.
Without oversight, operational risk grows.
๐๐๐๐๐๐ ๐๐๐:
Visibility and control are essential.
โธ Swipe
Learn PHP in 2026 (Yes, Really)
https://fagnerbrack.com/learn-php-in-2026-yes-really-bd567753dd84
#HackerNews #LearnPHP #PHP2026 #TechTrends #ProgrammingEducation #HackerNews
The complete IPv4 address space, mapped
#HackerNews #IPv4 #IPv4Mapping #AddressSpace #TechTrends #Networking
The OnlyFans Economy of American AI
https://leoveanu.com/2026-06-06-qwen3.7max/
#HackerNews #OnlyFans #Economy #AmericanAI #AIIndustry #DigitalEconomy #TechTrends
Netlify CTO Dana Lawson: Writing code is no longer the job
https://thenewstack.io/netlify-agent-experience-engineers/
#HackerNews #Netlify #DanaLawson #CodeFuture #TechTrends #EngineeringInsights
N-gated Hacker News
Hacker News
InfoQ
WanadevDigital
Halexo Ltd
SquaredTech