--- title: ChatGPT Work does scheduled tasks, but devs want a full... description: ChatGPT Work now supports scheduled cloud tasks that run without your laptop being on, connecting to Gmail, Google Drive, Slack, and other apps via plugins.... canonical: https://daily.dev/posts/chatgpt-work-does-scheduled-tasks-but-devs-want-a-full-cloud-pc-hggcftkha twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: ChatGPT Work does scheduled tasks, but devs want a full cloud PC | daily.dev og:description: ChatGPT Work now supports scheduled cloud tasks that run without your laptop being on, connecting to Gmail, Google Drive, Slack, and other apps via plugins.... og:url: https://daily.dev/posts/chatgpt-work-does-scheduled-tasks-but-devs-want-a-full-cloud-pc-hggcftkha og:image: https://api.daily.dev/og/posts/HgGCftkhA.png og:image:alt: ChatGPT Work does scheduled tasks, but devs want a full cloud PC 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. # ChatGPT Work does scheduled tasks, but devs want a full cloud PC **[Trends](https://daily.dev/sources/trends)** · 2 min read · 3 upvotes · 1 comments ## Summary ChatGPT Work now supports scheduled cloud tasks that run without your laptop being on, connecting to Gmail, Google Drive, Slack, and other apps via plugins. While useful for recurring automations, the feature has a key limitation: the cloud browser can't sign into accounts or reuse existing sessions, meaning every authenticated action must go through an official plugin integration. Developers are frustrated because what they actually want is a persistent cloud PC — a browser profile that stays logged in, preserves files and context between sessions, and treats plugins as a fallback rather than the only path. The gap between scheduled task automation and true web-wide agent action involves unsolved problems around session security, credential storage, and liability. ## Content OpenAI launched ChatGPT Work on July 9th, an agentic product that runs coding and knowledge tasks inside isolated cloud microVMs, connects to Gmail, Slack, Salesforce and others via MCP servers, handles browser automation, and supports scheduled recurring tasks. It hit 10 million users fast enough that OpenAI stopped announcing milestones. Greg Brockman's take: just ask it to do any recurring task. The community noticed the silence. "No 11M users yet?" became a pointed question, partly because each previous milestone had come with a quota reset for paid users. OpenAI's non-answer — "way past it, we stopped announcing at 10M" — gave direction but no definition, no time window, no cohort breakdown. For heavy users, every new million had started to look like a possible capacity event. Stopping the announcements without separating the reset policy from the growth celebration created exactly the ambiguity you'd expect. The deeper problem is that "user" is becoming the wrong unit. One person might supervise several parallel agents running for hours; another might have opened the product once and never trusted it with anything real. OpenAI's own research found 99th-percentile users accumulating roughly 71 hours of agent turns in a single day. A raw headcount hides whether growth is coming from developers, knowledge workers, enterprise seats, or dormant accounts reactivated by a free reset. On the product side, developers are already bumping into the gaps. Peter Yang's read: Work is one step away from the persistent cloud computer he actually wants. The browser can automate public sites but can't sign into accounts or reuse existing sessions. Scheduling works through the ChatGPT Work tab but apparently not from the Codex tab, which is the kind of friction that kills recurring workflows before they start. The ask is straightforward — a persistent browser profile, secure session reuse, a working environment that survives between tasks — and the team is presumably already on it. The 10M number is real. What it doesn't tell you is how many of those users delegated something meaningful, accepted the result, and came back the following week. That's the metric that would actually matter, and OpenAI isn't publishing it yet. ## Questions this post answers ### What are the current limitations of ChatGPT Work's cloud browser? ChatGPT Work's cloud browser cannot sign into accounts or reuse existing browser sessions, meaning it only works on public sites. Authenticated services still require plugins. What users actually want is a persistent cloud computer with a real browser profile that keeps running after closing the laptop — that capability is not yet available. _Developers evaluating AI agent tools for real workflows track gaps like these on daily.dev before committing to a stack._ ### What does OpenAI's ChatGPT Work actually do technically? ChatGPT Work runs scheduled tasks in isolated cloud microVMs, connects to external services like Gmail, Slack, and Salesforce via MCP servers, and handles browser automation through a managed Chrome instance. It launched on July 9th and is positioned as an agentic product for delegating multi-step tasks rather than a simple chat interface. _Teams building or adopting agentic workflows can follow ChatGPT Work developments alongside competing tools on daily.dev._ ### How do you measure meaningful retention for AI agent products instead of raw user counts? Raw user counts conflate heavy operators running parallel agents for hours with one-time visitors who never returned. A more meaningful metric is retained operators: users who delegated a real task, accepted the result, and came back the following week. OpenAI's own data found 99th-percentile users accumulating roughly 71 hours of agent turns in a single day, illustrating how misleading headcount alone can be. _Product engineers rethinking agentic product metrics will find the ongoing debate covered on daily.dev._ ## Community discussion Top comments from developers on daily.dev. **@trevorsuna** · 0 upvotes > A persistent cloud PC sounds simple until session security and recovery enter the picture. Plugins are narrower, but that constraint also limits the blast radius; the useful middle ground may be scoped browser identities with explicit expiry and audit logs. --- Tags: [#automation](https://daily.dev/tags/automation), [#ai-agents](https://daily.dev/tags/ai-agents), [#openai](https://daily.dev/tags/openai), [#chatgpt](https://daily.dev/tags/chatgpt) [View this post on daily.dev](https://daily.dev/posts/chatgpt-work-does-scheduled-tasks-but-devs-want-a-full-cloud-pc-hggcftkha) ```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":"ChatGPT Work does scheduled tasks, but devs want a full cloud PC","url":"https://daily.dev/posts/chatgpt-work-does-scheduled-tasks-but-devs-want-a-full-cloud-pc-hggcftkha","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/chatgpt-work-does-scheduled-tasks-but-devs-want-a-full-cloud-pc-hggcftkha"},"datePublished":"2026-08-02T03:04:04.952Z","dateModified":"2026-08-06T17:42:37.843Z","description":"ChatGPT Work now supports scheduled cloud tasks that run without your laptop being on, connecting to Gmail, Google Drive, Slack, and other apps via plugins....","isAccessibleForFree":true,"articleSection":"Trends","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":"Trends","logo":"https://media.daily.dev/image/upload/s--ZfSp3asX--/f_auto,q_auto/v1780996004/logos/trends?_a=BAMAMiWQ0","url":"https://daily.dev/sources/trends"},"commentCount":1,"discussionUrl":"https://daily.dev/posts/chatgpt-work-does-scheduled-tasks-but-devs-want-a-full-cloud-pc-hggcftkha","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":3},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":1}],"keywords":"automation,ai-agents,openai,chatgpt","timeRequired":"PT2M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Trends","item":"https://daily.dev/sources/trends"},{"@type":"ListItem","position":3,"name":"ChatGPT Work does scheduled tasks, but devs want a full cloud PC"}]} {"@context":"https://schema.org","@type":"WebPage","@id":"https://daily.dev/posts/chatgpt-work-does-scheduled-tasks-but-devs-want-a-full-cloud-pc-hggcftkha","comment":[{"@type":"Comment","text":"A persistent cloud PC sounds simple until session security and recovery enter the picture. Plugins are narrower, but that constraint also limits the blast radius; the useful middle ground may be scoped browser identities with explicit expiry and audit logs.","datePublished":"2026-08-02T04:04:21.782Z","url":"https://daily.dev/posts/HgGCftkhA#c-1f15CcuH6","author":{"@type":"Person","name":"Trevor Suna","url":"https://daily.dev/trevorsuna","image":"https://media.daily.dev/image/upload/s--dZ7gXxpp--/f_auto/v1784081551/avatars/avatar_EMoP47rpuw8DNjhp6R1b6?_a=BAMAMicg0"}}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/chatgpt-work-does-scheduled-tasks-but-devs-want-a-full-cloud-pc-hggcftkha#faq","mainEntity":[{"@type":"Question","name":"What are the current limitations of ChatGPT Work's cloud browser?","acceptedAnswer":{"@type":"Answer","text":"ChatGPT Work's cloud browser cannot sign into accounts or reuse existing browser sessions, meaning it only works on public sites. Authenticated services still require plugins. What users actually want is a persistent cloud computer with a real browser profile that keeps running after closing the laptop — that capability is not yet available. Developers evaluating AI agent tools for real workflows track gaps like these on daily.dev before committing to a stack."}},{"@type":"Question","name":"What does OpenAI's ChatGPT Work actually do technically?","acceptedAnswer":{"@type":"Answer","text":"ChatGPT Work runs scheduled tasks in isolated cloud microVMs, connects to external services like Gmail, Slack, and Salesforce via MCP servers, and handles browser automation through a managed Chrome instance. It launched on July 9th and is positioned as an agentic product for delegating multi-step tasks rather than a simple chat interface. Teams building or adopting agentic workflows can follow ChatGPT Work developments alongside competing tools on daily.dev."}},{"@type":"Question","name":"How do you measure meaningful retention for AI agent products instead of raw user counts?","acceptedAnswer":{"@type":"Answer","text":"Raw user counts conflate heavy operators running parallel agents for hours with one-time visitors who never returned. A more meaningful metric is retained operators: users who delegated a real task, accepted the result, and came back the following week. OpenAI's own data found 99th-percentile users accumulating roughly 71 hours of agent turns in a single day, illustrating how misleading headcount alone can be. Product engineers rethinking agentic product metrics will find the ongoing debate covered on daily.dev."}}]} ```