--- title: Why Agents Need Computers, Not Sandboxes | daily.dev description: A founder discusses the concept of building persistent 'computers for agents' instead of ephemeral sandboxes, arguing that AI agents inherently understand... canonical: https://daily.dev/posts/why-agents-need-computers-not-sandboxes-iep3qtj5g twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Why Agents Need Computers, Not Sandboxes | daily.dev og:description: A founder discusses the concept of building persistent 'computers for agents' instead of ephemeral sandboxes, arguing that AI agents inherently understand... og:url: https://daily.dev/posts/why-agents-need-computers-not-sandboxes-iep3qtj5g og:image: https://api.daily.dev/og/posts/Iep3qTj5G.png og:image:alt: Why Agents Need Computers, Not Sandboxes og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Why Agents Need Computers, Not Sandboxes **[Fly.io](https://daily.dev/sources/flydotio)** · 1 min read · 0 upvotes · 0 comments ## Summary A founder discusses the concept of building persistent 'computers for agents' instead of ephemeral sandboxes, arguing that AI agents inherently understand Linux computer primitives because LLMs were trained on massive amounts of internet text about Linux systems and servers. As agents run for hours or days rather than seconds, ephemeral sandboxes fail to provide the durability needed, making full, persistent Linux-based computers a better fit for how coding agents like Claude Code and Codex actually operate. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### Why do AI coding agents work better with a full Linux computer instead of an ephemeral sandbox? AI agents perform better with persistent Linux computers because LLMs were trained on vast amounts of internet text about Linux systems and servers, giving them native fluency in Linux command-line primitives like reading, writing, and searching files. Coding harnesses such as Claude Code and Codex run loops of these simple Linux commands, and as agents increasingly run for hours or days rather than seconds, ephemeral sandboxes lack the durability needed to support that. _Developers weighing sandbox versus persistent compute for AI agents can follow this debate on daily.dev._ ### Why are ephemeral sandboxes considered insufficient for long-running AI agent tasks? Ephemeral sandboxes are considered insufficient because they are not designed to persist for the hours or days that increasingly capable agents now take to complete tasks. As agent runtimes grow beyond seconds and minutes, a temporary sandbox environment cannot provide the durability and continuity required, making a persistent full-computer environment a better architectural fit. _Teams designing infrastructure for long-running agents can track this shift toward persistent compute on daily.dev._ ## Similar posts on daily.dev - [Code And Let Live](https://daily.dev/posts/code-and-let-live-haeiv3yfw) · Fly.io · 4 upvotes · 0 comments - [Development is about to outgrow the laptop](https://daily.dev/posts/development-is-about-to-outgrow-the-laptop-axfim0m5f) · Dev Interrupted · 0 upvotes · 0 comments - [Sandboxing AI Agents](https://daily.dev/posts/sandboxing-ai-agents-5lrrozvvp) · OctopusDeploy · 3 upvotes · 1 comments --- Tags: [#linux](https://daily.dev/tags/linux), [#llm](https://daily.dev/tags/llm), [#infrastructure](https://daily.dev/tags/infrastructure), [#ai-agents](https://daily.dev/tags/ai-agents) [View this post on daily.dev](https://daily.dev/posts/why-agents-need-computers-not-sandboxes-iep3qtj5g) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Why Agents Need Computers, Not Sandboxes","url":"https://daily.dev/posts/why-agents-need-computers-not-sandboxes-iep3qtj5g","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/why-agents-need-computers-not-sandboxes-iep3qtj5g"},"datePublished":"2026-08-18T21:05:54.797Z","dateModified":"2026-08-20T11:55:49.430Z","description":"A founder discusses the concept of building persistent 'computers for agents' instead of ephemeral sandboxes, arguing that AI agents inherently understand...","image":"https://i.ytimg.com/vi/P17eNVUlyR8/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/P17eNVUlyR8/sddefault.jpg","isAccessibleForFree":true,"articleSection":"Fly.io","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Fly.io","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/deb71e3b6f964388876ecc82c26c8f6c","url":"https://daily.dev/sources/flydotio"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/why-agents-need-computers-not-sandboxes-iep3qtj5g","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"linux,llm,infrastructure,ai-agents","timeRequired":"PT1M","video":{"@type":"VideoObject","name":"Why Agents Need Computers, Not Sandboxes","description":"A founder discusses the concept of building persistent 'computers for agents' instead of ephemeral sandboxes, arguing that AI agents inherently understand...","thumbnailUrl":"https://i.ytimg.com/vi/P17eNVUlyR8/sddefault.jpg","uploadDate":"2026-08-18T21:05:54.797Z","duration":"PT1M","url":"https://api.daily.dev/r/Iep3qTj5G","embedUrl":"https://www.youtube.com/embed/P17eNVUlyR8"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Fly.io","item":"https://daily.dev/sources/flydotio"},{"@type":"ListItem","position":3,"name":"Why Agents Need Computers, Not Sandboxes"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/why-agents-need-computers-not-sandboxes-iep3qtj5g#faq","mainEntity":[{"@type":"Question","name":"Why do AI coding agents work better with a full Linux computer instead of an ephemeral sandbox?","acceptedAnswer":{"@type":"Answer","text":"AI agents perform better with persistent Linux computers because LLMs were trained on vast amounts of internet text about Linux systems and servers, giving them native fluency in Linux command-line primitives like reading, writing, and searching files. Coding harnesses such as Claude Code and Codex run loops of these simple Linux commands, and as agents increasingly run for hours or days rather than seconds, ephemeral sandboxes lack the durability needed to support that. Developers weighing sandbox versus persistent compute for AI agents can follow this debate on daily.dev."}},{"@type":"Question","name":"Why are ephemeral sandboxes considered insufficient for long-running AI agent tasks?","acceptedAnswer":{"@type":"Answer","text":"Ephemeral sandboxes are considered insufficient because they are not designed to persist for the hours or days that increasingly capable agents now take to complete tasks. As agent runtimes grow beyond seconds and minutes, a temporary sandbox environment cannot provide the durability and continuity required, making a persistent full-computer environment a better architectural fit. Teams designing infrastructure for long-running agents can track this shift toward persistent compute on daily.dev."}}]} ```