How to Use Claude's Massive New Upgrades
What this covers
Major theme: Claude ecosystem rapidly expanded with powerful model upgrades and a shift toward persistent, context-rich agent workflows. Key features: remote control for local terminals, Dispatch persistent cowork orchestration, Channels for external events, and direct computer control via mouse, keyboard and screen. Consequences discussed include reimagined workflows, broad automation potential for legacy and modern apps, and heightened security and access-control challenges.
The AI Daily Brief helps you understand the most important news and discussions in AI. Subscribe to the podcast version of The AI Daily Brief wherever you listen: https://pod.link/1680633614 Get it ad free at http://patreon.com/aidailybrief Learn more about the show https://aidailybrief.ai/
Source description (no synthesized summary yet).
Anthropic has released a comprehensive suite of features over the past month that fundamentally transforms how users interact with Claude by enabling persistent, context-maintaining, always-on work orchestration accessible from any device, positioning Claude as a true autonomous execution layer rather than just a tool.
- Computer use capability allows Claude to control mouse, keyboard, and screen to automate tasks in any application without setup
- Dispatch, remote control, and channels create multiple modalities for asynchronous task initiation and monitoring from mobile devices
- Scheduled tasks enable work to run continuously without user presence, treating Claude as a delegated agent rather than a tool to operate
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Computer use capability allows Claude to control mouse, keyboard, and screen to interact with any application by clicking, navigating, and typing, enabling task completion in non-API-enabled legacy software without requiring setup or pre-built connectors.
“Today we're releasing a feature that allows Claude to control your computer, mouse, keyboard, and screen, giving it the ability to use any app. When Claude doesn't have access to the tools it needs, it will point, click, and navigate what's on your screen to perform the task itself. It can open files, use the browser, and run dev tools automatically with no setup required.”
Deployment of agentic systems at scale will require solving multiple classes of unsolved security problems including identity (whether agents act on behalf of users or have independent identities with limited access rights), behavior monitoring (historical activity volume is no longer a reliable signal of compromise), and prompt injection prevention to prevent agents from being manipulated into risky actions.
“To be sure, there are going to be various hurdles around security, permissions, and access control, identity challenges, and more. For instance, should the agent always act on behalf of the user, or should they have their own identity and limited set of access rights? How do you triage security events when historically volume of activity on a system is no longer a reliable signal of a security issue? How do you ensure the agent isn't going rogue or getting prompt injected to do something risky?”
Dispatch in Claude co-work enables a single persistent conversation thread with Claude that retains context across multiple spawned sessions (development in Claude Code, knowledge work in co-work), allowing users to message Claude from mobile devices and receive notifications when tasks complete or require approval.
“We're shipping a new feature in Claude co-work as a research preview that I'm excited about, dispatch. One persistent conversation with Claude that runs on your computer. Message it from your phone, come back to finished work.”
Claude Code channels enable push-based event integration (rather than pull-based tool calls) where external systems like monitoring alerts, CI failures, or webhook payloads can push events into a running Claude Code session through MCP servers on platforms like Telegram or Discord, allowing Claude to react to external stimuli without user intervention.
“We just released Claude code channels, which allows you to control your Claude code session through select MCPs, starting with Telegram and Discord. A channel is an MCP server that pushes events into your running Claude code session, so Claude can react to things that happen while you're not at the terminal.”
Scheduled tasks were added to co-work first (end of February), then to Claude Code desktop (week and a half later as local scheduling), then as recurring cloud-based tasks (couple weeks after), demonstrating a pattern of feature progression from simple to complex architectures.
“At the end of February, Co-work got scheduled tasks. Think a morning briefing, weekly spreadsheet updates, or Friday team presentations. Then about a week and a half later, we got local scheduled tasks in Claude Code desktop. Turic said that his favorite use case is to ask it to check error logs every few hours and create PRs for any actionable errors. Now, because these tasks are local, they run as long as your computer is awake. But then a couple weeks after that, we got recurring cloud-based tasks. or repos a schedule and a prompt. Claude runs it via cloud infra on your schedule so you don't need to keep Claude Code running on your local machine.”
Computer use announcement tweet reached 40 million views and 62,000 bookmarks within 16 hours of posting, indicating exceptional viral engagement and high market interest in the capability.
“About 16 hours after that tweet went live, 40 million people have viewed it. 62,000 people have bookmarked it, and the chatter is overwhelming.”
Claude Code's code review feature deploys multiple agent instances to hunt for bugs, treating bug detection as a team problem rather than individual agent task, improving code quality through parallel agent collaboration.
“Also in March, the Claude Code team launched code review, where as they put it, Claude dispatches a team of agents to hunt for bugs.”
Remote control enables Claude Code sessions started locally on a machine to be controlled remotely from mobile devices while maintaining full access to the local file system, MCP servers, tools, and product configurations, with code execution happening locally rather than in the cloud.
“The way that remote control works is that you start your task in your CloudCode terminal session, and then you can pick it up and continue working from your phone. Now, there's nothing happening in the cloud here. Basically, when you've started that remote control session on your machine, Cloud is going to continue to run locally that entire time. That gives remote control the ability to use your full environment, including your file system, MCP servers, tools, and product configurations”
Skills can now be saved in Excel and PowerPoint add-ins, allowing teams to standardize workflows (e.g., variance analysis, client deck building) and enable other users to execute these skills in one click from their sidebar.
“Skills are also now available inside the Excel and PowerPoint add-ins, meaning that when your team has some standard workflow, like running a variance analysis or building a client deck, you can save that as a skill. Other people in your organization can then run it in one click from their own sidebar.”
Remote control represents not merely a productivity feature but a fundamental shift in the mental model of AI tool interaction, moving from 'a tool you operate' to 'something you delegate to and check in with,' requiring lived experience rather than theoretical understanding to fully grasp.
“Claude code remote control just clicked for me. You kick off a task in the terminal, then pick it up from your phone on a walk. That's not a productivity feature. That's a relationship shift. You stop thinking of it as a tool you operate and start thinking of it as something you delegate to and check in with. Different mental model entirely.”
Computer use and code writing capabilities represent ultimate primitives for agent systems because most knowledge work requires hopping between multiple applications and handling diverse data sets, and agents need system traversal ability to automate real enterprise work beyond single-system optimization.
“Computer use and the ability to write and run code on the fly are the ultimate primitives for agents to be able to take on more and more tasks and knowledge work. Most work requires hopping between multiple applications and working with broad sets of data in a workflow, and agents will need to be able to traverse these systems to be able to effectively automate any real work in the enterprise.”
Anthropic shipped an extensive suite of features over the last month in response to competitive pressure from Open Claude's popularity, collectively representing a strategic shift toward making Claude an always-on, context-maintaining, persistent orchestration experience.
“Boy, it sure seems like Anthropic did not like everyone talking about open Claude instead of Claude Code because over the last month they have shipped an absolute boatload of features”
Local scheduled tasks run continuously as long as the user's computer remains awake, creating a practical constraint that differs from cloud-based persistent execution.
“Now, because these tasks are local, they run as long as your computer is awake.”
The 1 million token context window became generally available for both Claude Opus and Claude Sonnet during March, enabling significantly more complex analysis within a single context window.
“This past month, the 1 million token context window actually became available generally for both Claude Opus and Claude Sonnet.”
The suite of Claude features enables a fundamentally different software development process for AI-powered coding teams, with implications for how development velocity and team productivity can be structured when code generation and testing can be asynchronously automated.
“The ability of the Claude team to learn from things like open Claude and implement features like this on a daily basis is a very strong argument that for AI-powered coding teams, a very different software development process is possible with large strategic implications.”
Aaron Levie identifies software integration and API readiness as critical infrastructure challenges, noting that incumbents will face disruption if software cannot operate headlessly while companies with seamless API exposure and business models supporting it will benefit disproportionately.
“Then, there's also lots of work needed to ensure software is set up to enable agents to operate with their tools in a headless fashion. This will be an uncomfortable reality for some incumbents and equally a welcome one for tools that historically have operated seamlessly via APIs and have business models to support this.”
In a 48-hour real-world test, Pavel Huryn completed approximately 25 minutes of total directing time across multiple dispatch tasks while Claude executed in parallel for 3+ hours of equivalent work, demonstrating significant leverage in asynchronous work management.
“Ultimately, he writes, 'The actual direction time across all of these gaps was maybe 25 minutes total. The Claude execution running in parallel was 3 plus hours of work.'”
A developer built a custom orchestration system spawning Claude Code sessions across Docker, VMs, pods, and local MacBook in just 4 days after Channels launched, developing an iOS/iPadOS app for remote session control and identifying numerous additional customization possibilities.
“My goal was twofold. Build a custom orchestration system to spawn Claude Code sessions anywhere, Docker, VMs, pods, or locally on my MacBook, and then create a custom app across macOS, iOS, and iPadOS to control these agents on the go. Fast forward to today, just 4 days after channels dropped, and the iOS app is already up and running. I have so many more ideas on how to tweak and expand this to fit my exact workflow.”
Channels are particularly suited for technically sophisticated users who want hackable, customizable integrations, providing an alternative to dispatch which targets less technical users, with Tarek positioning channels as the choice for developers who want infrastructure control.
“Channels is more focused on devs who want something hackable.”
Computer use is particularly valuable for enterprises that depend on legacy software not designed for API integration, where traditional automation approaches cannot reach, making it a competitive advantage for organizations without modern tech stacks.
“This is another domino to fall. Computer use of arbitrary apps, not just your browser. This is a big deal for a lot of corporates who have custom crappy apps from 20 or 30 years ago.”
Peter Levels assessed Remote Control as 'extremely nice' and comparable favorably to SSH sessions, noting the ability to edit on macOS or iOS in the Cloud app on a production server from anywhere, suggesting it significantly reduces friction for distributed work.
“Prominent solopreneur Peter Levels wrote, 'Cloud remote control is extremely nice. Can edit on macOS or iOS in Cloud app on my production server from anywhere.' He basically compared it favorably to an SSH session, which would be another more technically complex way to log into your local device to control it while on the go.”
The primary takeaway for users regarding the feature releases is pragmatic necessity: they must test all features individually to understand how they fit their specific use cases, as each offers different trade-offs and capabilities.
“This is kind of why you're going to want to go hack at all of these to understand how they fit your own use cases”
Early examples of computer use were relatively simple (file grabbing, liking posts, opening apps) but represent a proof-of-concept that the capability works, with complexity and sophistication expected to increase as users understand the system.
“In terms of examples, a lot of people started fairly simply. Gavin Purcell writes, 'Okay, this is pretty cool. Just had it grab a photo off my desktop from afar. Dumb and simple, yes, but done in natural language with no weird setup.' Daniel Son writes, 'I don't like posting about anything I haven't tried myself, so I saw the announcement that Claude dispatched now controls your computer, and I simply asked it to open Twitter and like a Claude post. Done. This is getting more interesting and more dangerous by the minute.'”
Roman Mirzoyan fixed two bugs and released an app update to the App Store using remote control without touching his laptop, demonstrating that production-grade software operations can be performed through mobile interfaces.
“Roman Mirzoyan writes, 'Yesterday I fixed two bugs and then released an app update to the App Store without touching my laptop and having a walk for half a day.'”
Evan Malek assessed Dispatch as covering 90% of what he was attempting with Open Claude while feeling 'far less likely to upload my entire drive to a malware site', noting advantages in ease, stability, safety, and existing connectors (Gmail, browsers) and citing missing multi-session support.
“Even Malek also had a positive experience with dispatch. He writes, 'After using it a bit, Claude co-work dispatch covers 90% of what I was trying to use Open Claude for, but feels far less likely to upload my entire drive to a malware site.' 'What I like better,' he writes, 'easy, much more stable and safe, existing connectors mean better integration with Gmail, browsers, etc. Very good tool use. What's missing for me? The ability to invite Claude to any channel, the heartbeat and proactivity, and the multiple sessions.'”
Claude app now supports interactive charts and diagrams, enabling Claude to generate visualizations and graphical outputs beyond text-based responses.
“Meanwhile, over in the main Claude app, Claude can now build interactive charts and diagrams”
A plugin marketplace for enterprise customers enables organizations to extend Claude with custom integrations and tools, supporting tailored workflows without requiring development overhead from Anthropic.
“We also got a whole plugin marketplace for enterprise customers.”
Co-work in Claude now supports project abstractions, which addresses a significant usability gap where previous versions lacked the concept of organizing work around distinct projects, improving day-to-day functionality.
“One big complaint that lots of folks have had about Cohere Work is that there wasn't a conception of a project. Now there is, which will make a huge difference for day-to-day functionality.”
Memory and connectors are now available on Claude's free plan, lowering the barrier for new users to access context persistence and third-party integrations without paid subscription.
“For my free users out there, memory and connectors are now both available on the free plan”
Peter Steinberger's hire by OpenAI from Open Claude's founding prompted initial speculation that Anthropic had made a strategic error, but subsequent events (Anthropic's rapid feature releases) suggest that 'let him cook' advocates were vindicated in their assessment.
“When Open Claude's founder Peter Steinberger was hired by OpenAI, some people jumped in to say that Anthropic had made a big goof by not bringing him over there. Others' response was a little bit closer to, 'I don't know, man. Let's let him cook and see what happens.' And certainly you have to think that the people saying let them cook are feeling pretty vindicated right now.”
Daniel Son tested computer use to open Twitter and like a Claude post from mobile, demonstrating that the capability works for arbitrary app interactions, noting this increases both capability and potential dangers.
“Daniel Son writes, "I don't like posting about anything I haven't tried myself, so I saw the announcement that Claude dispatched now controls your computer, and I simply asked it to open Twitter and like a Claude post. Done. This is getting more interesting and more dangerous by the minute."”
Noah Weiman (Claude Code PM) joked that cloud-based recurring tasks were motivated by his inability to set up a Mac mini rather than deliberate feature design, suggesting some technical decisions emerge from practical necessity rather than strategic planning.
“He said that so far the Claude Code team had found it useful for things like sweeping through open PRs, building features from approved issues, analyzing CI failures overnight, and syncing docs based on newly merged PRs. And if you're wondering if this had to do with open claw, in another tweet Noah said, 'Some might say I built this because I couldn't figure out how to set up my Mac mini.'”
The month of releases represents not the end but the middle of Anthropic's feature development, with additional second-tier and quality-of-life announcements still pending beyond what was covered in the main feature categories.
“Now, we're already going pretty long, but I want to point out that this is not the end of what has been pushed over the last month. Running quickly through some of these second-tier and quality-of-life type of announcements”