Energy startup Arbor Energy on Wednesday said it had sold up to 5 gigawatts worth of its modular turbines to GridMarket, a company that helps arrange power projects for data centers and industrial users. 

“Everyone wants more power. They wanted it yesterday,” Brad Hartwig, co-founder and CEO of Arbor, told TechCrunch. “The time frames are compressing and the scale is getting larger.”

Arbor’s Halcyon turbines are based on rocket turbomachinery, high-performance engine technology originally developed for spaceflight, and its first commercial turbines will be 3D printed and capable of generating 25 megawatts each. GridMarket’s order, if fully fulfilled, represents 200 units.

Neither company disclosed the exact price of the deal, though Hartwig said that Arbor has seen a “willingness to pay of upwards of $100 per megawatt-hour.” A person familiar with the deal told TechCrunch that the total is in the single-digit billions of dollars.

The startup plans to connect its first turbine to the grid in 2028 and ramp production through 2030, at which point it hopes to deliver more than 100 turbines annually. The goal, Hartwig said, is to eventually produce enough for 10 gigawatts of new capacity every year.

Arbor’s initial designs intended for Halcyon to subsist on a vegetarian diet — the power plant would ingest organic material like crop waste and wood scraps from farms and timber operations, which would be turned into syngas — a combustible gas mixture — and burned in the presence of pure oxygen. The result would be pure CO₂, which could be captured and stored underground.

Under that process, each Halcyon turbine would generate carbon negative power. The organic matter it consumes would otherwise have decayed, releasing methane and carbon dioxide into the atmosphere.

Since then, Arbor has modified Halcyon to accept natural gas in addition to biomass — making it, in effect, more of an omnivore. The process otherwise remains the same, meaning the CO₂ that emerges can still be sequestered. 

Because it’s using natural gas, it wouldn’t be carbon negative in that configuration. In fact, because methane leaks from pipes and valves throughout the supply chain, Halcyon turbines running on fossil fuel will still produce some greenhouse gas emissions while also fostering continued demand for natural gas. Hartwig said that the company is working with low-leak natural gas suppliers, and that it’s “economically a benefit to sequester that CO₂.”

“We see a long-term path to less than 10 grams of CO₂ per kilowatt-hour,” Hartwig said. That’s significantly lower than typical natural gas power plants without carbon capture, which release about 400 grams of CO₂ per kilowatt hour.

Arbor hasn’t abandoned its biomass-powered projects, and the sale to GridMarket isn’t restricted to one specific fuel. However, other announced deals built around biomass are considerably smaller than the one signed with GridMarket.

Like many energy startups, Arbor has gotten a meaningful boost from the data center boom. Makers of traditional gas turbines were caught flat-footed, and given the volatility of such markets in the past, they’ve been reticent to significantly increase production. Hartwig said that they’d be hard-pressed to quickly ramp production, even if they wanted to. 

“Those supply chains largely all get bottlenecked by blades and vanes for traditional turbines. Those are fairly inelastic supply chains, both in how artisanal the production method is — doing directionally solidified, single-crystal turbine blades — as well as very specialized labor, the workforce behind it,” he said. “If you were to get in line for a turbine today, you’d be waiting until 2032.”

Arbor is betting that its machined and 3D printed parts will help it get to market quicker. “People want power in the next few years and they want a lot of it,” Hartwig said.

Kali Linux 2026.1, the first release of the year, is now available for download, featuring 8 new tools, a theme refresh, and a new BackTrack mode for Kali-Undercover.

Kali is a Linux distribution designed for ethical hackers and cybersecurity professionals, with tools for red teaming, penetration testing, network research, and security assessments.

This distribution is available as a live environment or an installable operating system, and it supports a wide range of hardware, including Raspberry Pi devices and compatible Android phones via Kali NetHunter.

New tools added to Kali Linux 2026.1

In this release, the Kali Team has added 25 new packages, updated 183 others, and upgraded the Kali kernel to 6.18.

It has also added new tools to the network repositories, including:

• AdaptixC2 – Extensible post-exploitation and adversarial emulation framework
• Atomic-Operator – Execute Atomic Red Team tests across multiple operating system environments
• Fluxion – Security auditing and social-engineering research tool
• GEF – Modern experience for GDB with advanced debugging capabilities
• MetasploitMCP – MCP server for Metasploit
• SSTImap – Automatic SSTI detection tool with interactive interface
• WPProbe – Fast WordPress plugin enumeration tool
• XSStrike – Advanced XSS scanner

Annual theme refresh

With this year’s first release, the Kali Team introduced a theme update that adds new wallpapers, tweaks the graphical installer interface, and improves the boot and login experience.

“As with previous 20xx.1 releases, this major update brings our annual theme refresh, a long-standing tradition that keeps the Kali Linux interface as modern and innovative,” the Kali devs said.

“This year’s release unveils a brand-new theme from the moment you boot. Everything from the boot menu, installer to the login display, and a fresh set of desktop wallpapers.”

BackTrack mode for Kali-Undercover

This version also adds a new “BackTrack mode” to the distro’s Kali Undercover, a feature previously that may also be used to make the Kali desktop look like a default Windows 10 installation.

With the new mode, users can quickly switch to a theme closely resembling Kali’s predecessor, BackTrack Linux.

“This mode transforms the desktop to recreate the look and feel of BackTrack 5, with the same wallpaper, colors, and window themes,” the announcement reads.

“You can run it directly from the menu or by running kali-undercover –backtrack in the terminal. You can switch back to the default Kali desktop (or not) by running it again.”

The Kali Team also made improvements to the Kali NetHunter app, including a HID permission check and fixes for the WPS scan bug and the back button issue.

How to get Kali Linux 2026.1

To start using Kali Linux 2026.1, you can upgrade your existing installation, select a platform, or directly download ISO images for new installs and live distributions.

Kali users who want to update from a previous version can use the following commands to upgrade to the latest version.

echo "deb http://http.kali.org/kali kali-rolling main contrib non-free non-free-firmware" | sudo tee /etc/apt/sources.list
sudo apt update && sudo apt -y full-upgrade
cp -vrbi /etc/skel/. ~/
[ -f /var/run/reboot-required ] && sudo reboot -f

If you’re using Kali on Windows Subsystem for Linux, you should consider upgrading to WSL 2 for better support of graphical applications. To check your WSL version, run wsl -l -v in a Windows command prompt.

After upgrading, you can check whether the process was successful using the following command: grep VERSION /etc/os-release.

You can view the complete changelog for Kali Linux 2026.1 on Kali’s website.

The new jobs are part of Newcode’s international expansion, following a seed funding round of more than €5m.

Norwegian software company Newcode is continuing its European expansion with plans to hire 30 employees as it develops a presence in Dublin, Ireland. The move is part of Newcode’s plans for international growth, following a seed-funding round of more than €5.7m, which was backed by Alliance VC, The LegalTech Fund and other major investors.

Founded in 2021, Newcode is an Oslo-headquartered organisation with an additional presence across Europe and the US. The company enables firms and enterprises to run intelligent, auditable and context-aware workflows across their legal operations. Currently the company is pitching its agentic AI platform as a digital co-worker for legal teams and processes, as an agent that sits between a firm’s systems, data and applications. 

Commenting on the jobs announcement, Maged Helmy, Newcode’s co-founder and CEO, said, “As we expand our international footprint, we’re excited to establish our presence in Ireland starting with new 30 hires planned in the first year as we scale and invest in one of Europe’s most dynamic technology and professional services markets. 

“With operations already across Denmark, Oslo, Sweden, Finland, Palo Alto, and New York, Dublin is the natural next step in our growth.”

This week, also in Norway, Lace Lithography, a Microsoft-backed chipmaking equipment start-up, announced the raising of $40m to be used in the advancement of its unique semiconductor technologies. Lace’s engineers have developed a novel system that uses a helium atom beam over a light based system, to enable the design of chips that are 10 times as small as what is currently possible. 

The organisation’s CEO Bodil Holst explained Lace’s advancements could allow chipmakers to print silicon wafers at what could be considered “ultimately atomic resolution”. 

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

Last month, Iran’s Tehran Times posted what appeared to be damning satellite proof: a before-and-after image of “American radar,” supposedly “completely destroyed.”

It wasn’t. The image was an AI-manipulated version of a year-old Google Earth shot from Bahrain—wrong location, wrong timeline, fabricated damage. Open source intelligence researchers debunked it within hours matching it to older satellite imagery and identifying identical visual artifacts, down to cars frozen in the same positions.

A small act of disinformation, quickly debunked. But it pointed to a challenge that becomes more difficult during active conflict: The satellite infrastructure that journalists, analysts, pilots, and governments rely on to see conflict clearly in the Gulf is itself becoming contested terrain—delayed, spoofed, withheld, or simply controlled by actors whose interests don’t always align with public access.

The escalation follows rising tensions between the US, Israel, and Iran, with missile and drone activity crossing Gulf airspace and regional infrastructure—including satellites and navigation systems—entering into the conflict.

No Longer Neutral Infrastructure

When satellite data becomes unreliable, control over it becomes a central question.

In the Gulf, satellite infrastructure is largely run by state-backed operators. These rely on geostationary satellites—positioned high above the equator—which are used for activities such as broadcasting, communication and weather forecasting.

In the United Arab Emirates, that includes Space42 for secure communications and Earth observation. Saudi-led Arabsat handles broadcasting and broadband, while Qatar’s Es’hailSat supports regional connectivity. All operate under close government oversight.

Iran is building a parallel system. Its satellites, including Paya (also known as Tolou-3), are part of a broader push to expand surveillance capabilities independently of Western infrastructure. The high-resolution Earth observation satellite was launched from Russia’s Vostochny Cosmodrome.

The market around that infrastructure is growing fast. The Middle East satellite communications sector is valued at more than $4 billion and projected to reach $5.64 billion by 2031, according to one estimate, driven largely by airborne connectivity linked to both commercial aviation and defense demand. Maritime platforms already account for nearly a third of regional revenue.

Access Is the New Bottleneck

Commercial low-Earth orbit fleets like Planet Labs and Maxar operate differently from government-owned systems—and access is the main constraint. Governments receive priority tasking, while newsrooms and NGOs rely on paid subscriptions.

On March 11, Planet Labs announced it would extend delays on imagery of the Middle East by two weeks. The company denied the decision came from any government request, stating instead that it was to “ensure our imagery is not tactically leveraged by adversarial actors to target allied and NATO-partner personnel and civilians.”

Maryam Ishani Thompson, an open source intelligence (OSINT) reporter, tells WIRED Middle East that “the loss of Planet Labs is so harsh because we were getting a fast refresh rate. Even if we turn to Chinese satellites, we don’t get that speed.”

Chinese platforms like MizarVision, a Shanghai-based open source geospatial intelligence provider, have seen increased use since the delays—part of a broader shift in who controls the imagery pipeline. Russia and China are also increasingly sharing satellite access with Iran, meaning the companies that once set the terms of what the world could see are no longer the only ones with eyes on the Gulf.

If You Can’t Verify, You Can’t Challenge the Narrative

Operationally, the consequences are immediate.

Ishani’s verification process depends on historical reference points. The static nature of the Tehran Times image—with cars in identical positions across both frames—was detectable precisely because journalists had recent imagery to compare against. Remove that baseline, and the same image becomes harder to debunk.

“In that opaque space,” Ishani says. “Iran is producing its own false narrative. If we can’t document it and fact-check it, they can continue to create a narrative and sell it to their people.”

Victoria Samson, chief director of space security and stability at nonprofit Secure World Foundation, says that, for most commercially and privately owned satellite companies, the US government is one of their largest customer—creating “a reluctance to upset the US government.”

Ring has today announced a spec bump to its battery-powered video doorbells for all those folks who can’t wire their units to power. The flagship Battery Doorbell Pro (2nd gen) gets 4K video, with 10x zoom and the promise of far longer time between recharges than the previous model. At the same time, it’s bringing 2K imaging to its lower-end battery doorbells, the Battery Doorbell Plus and Battery Doorbell (2nd gen). The former, as fitting its higher price, gets a quick-release battery pack, while both models get 2K video and 6x zoom. Naturally, these features are already available on Ring’s wired products, the bulk of which were announced back in September 2025.

The company is also aware that swapping out batteries isn’t ideal if you really need a doorbell to work all of the time. That’s why it’s also launching a new Solar Charger which integrates into the mount, keeping your doorbell running for longer between trips to the wall outlet. There’s also a bigger Solar Panel, which pumps out more juice than its smaller sibling, and can be mounted in a wider variety of places. All of the above are available to pre-order from today, and are priced as follows: Pro ($250), Plus ($180), Battery Doorbell ($100), Solar Charger ($50), Solar Panel ($60).

Production will be shifted to its regional facilities in Malaysia and Vietnam

[Editor’s Note: This article has been updated with the latest developments on APBS’ job cuts and Tuas plant changes.]

Tiger Beer brewer Asia Pacific Breweries Singapore (APBS) is cutting jobs and scaling down its brewing operations in Singapore, according to media reports from The Business Times and Channel NewsAsia today (Mar 24).

The brewer plans to progressively phase down large-scale brewing at its Tuas plant by the end of 2027, with around 130 roles affected as production is shifted to regional facilities in Malaysia and Vietnam. Over time, the Tuas site will be redeveloped to support regional logistics and innovation activities, including a pilot brewery.

The operational changes in Singapore will be implemented “progressively.” APBS said it will work with the Food, Drinks and Allied Workers Union (FDAWU) to support affected employees with severance, reskilling, outplacement services and well-being resources.

The retrenchments aren’t coming out of nowhere. APBS last restructured in late 2023, cutting 33 jobs and giving affected staff severance, bonuses, and annual wage supplements.

Globally, parent company Heineken also flagged more cuts earlier this year, saying 5,000 to 6,000 jobs could go over the next two years as market conditions tighten. Singapore serves as its Asia-Pacific headquarters.

The move follows a 2.8% drop in consolidated beer volumes in 2025, with Europe and the Americas seeing declines of 4% and 3.6%, respectively. The two markets account for 68.2% of total beer volumes.

Industry experts note that declining alcohol consumption among younger consumers is reshaping the market. Many young adults are drinking later, in smaller quantities, or not at all, opting for experiences over intoxication.

That said, Asia-Pacific beer volumes still rose slightly by 0.4% to 4.6 billion litres.

APBS, formerly known as Malayan Breweries, launched Tiger Beer in 1932 through a partnership between Heineken and Fraser and Neave (F&N). The company rebranded as APBS in 1990, and Heineken acquired F&N’s stake for S$5.6 billion in 2012.

Vulcan Post has reached out to APBS for further information.

• Read more stories we’ve written on the latest job trends here.

Featured Image Credit: MR. AEKALAK CHIAMCHAROEN via Shutterstock.com/ Asia Pacific Breweries Singapore

UK RRP £2699 £4099 £2399 £1199 £1599 £1399 £1099 £1199 – £2399 £2199 USA RRP $2699 $3799 $1799 $1099 – – – Unavailable – – – EU RRP €3199 – €2099 €1239 – – – – – – – CA RRP – CA$3999 CA$2399 Unavailable – – – Unavailable – – – AUD RRP AU$4007 AU$6299 AU$3495 Unavailable – – – Unavailable – – – Manufacturer LG Samsung Sony Samsung Hisense LG Panasonic Philips Samsung TCL Sony Quiet Mark Accredited – – – – – – – – – – – Screen Size 64.5 inches 74.5 inches 54.6 mm 55 inches 64.5 inches 41.5 inches 47.6 inches 54.6 inches 64.5 inches 97.5 inches 54.6 inches Size (Dimensions) 1441 x 230 x 880 MM 1680 x 319.8 x 992.7 MM 1227 x 327 x 780 MM 1227.6 x 253 x 768 MM 1449 x 295 x 899 MM 932 x 170 x 577 MM x x INCHES x x INCHES x x MM 2180 x 420 x 1285 MM 1223 x 248 x 786 MM Size (Dimensions without stand) 826 x 1441 x 45 MM 958.1 x 1680 x 38.5 MM 712 x 1227 x 53 MM 706.2 x 1227.6 x 47.2 MM 838 x 1449 x 77 MM 540 x 932 x 41.1 MM x x INCHES 708 x 1228 x 58 MM x x MM 1247 x 2180 x 64 MM 706 x 1223 x 37 MM Weight 16.6 KG 35 KG 18.8 KG 19.8 KG 20.8 KG 10.1 KG – 17.2 KG 21.2 KG 54.6 KG 18 KG ASIN – – B0BX449WWF B0CYBPLQY8 B0CYQ92K8C B0DYQMWSKG – B0F24VHMK4 – – B0CZTZTQXJ Operating System webOS Tizen Google TV Tizen OS VIDAA 7.6 webOS 25 Fire TV Titan OS Tizen Google TV Google TV Release Date 2024 2025 2023 2024 2024 2025 2025 2025 2025 2025 2024 First Reviewed Date – – – 30/06/2024 – – – – – – – Model Number OLED65C46LA QE75QN900F XR-55A80L Samsung QE55Q80D 65U7NQTUK – – 55OLED760/12 – – K55XR80 Model Variants – – – 50Q80D, 65Q80D, 75Q80D, 85Q80D 65U7N – – – – – – Resolution 3840 x 2160 7680 x 4320 3840 x 2160 3840 x 2160 3840 x 2160 3840 x 2160 3840 x 2160 3840 x 2160 3840 x 2160 3840 x 2160 3840 x 2160 HDR Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Types of HDR HDR10, HLG, Dolby Vision HDR10, HLG, HDR10+ Adaptive HDR10, HLG, Dolby Vision HDR10, HLG, HDR10+ HDR10, HLG, Dolby Vision IQ, HDR10+ HDR10, HLG, Dolby Vision HDR10, HLG, Dolby Vision IQ, HDR10+ Adaptive HDR10, HLG, Dolby Vision, HDR10+ Adaptive HDR10, HLG, HDR10+ Adaptive HDR10, HLG, Dolby Vision, Dolby Vision IQ, HDR10+, HDR10+ Adaptive HDR10, HLG, Dolby Vision Refresh Rate TVs 40 – 144 Hz 48 – 165 Hz 40 – 120 Hz 24 – 120 Hz 40 – 144 Hz 48 – 144 Hz 48 – 144 Hz 48 – 120 Hz 40 – 144 Hz 48 – 144 Hz 40 – 120 Hz Ports Four HDMI 2.1, three USB inputs, LAN connector, digital audio output, satellite, RF Four HDMI 2.1, 2 x USB, Ethernet, RF input, optical digital audio output Four HDMI, digital optical out, two USB ports, composite video input, Ethernet, two satellite, terrestrial inputs Three USBs, Four HDMIs, optical digital audio output, Ethernet port, RF port Four HDMI, USB 2.0, USB 3.0, Ethernet, two RF inputs, CI+ 2.0 slot, digital optical out, 3.5mm audio output, sub-out, headphone out. AV composite input Four HDMI 2.1, three USB, ethernet, optical digital out, CI+, two RF tuners Four HDMI, digital audio output, Three USB ports, Ethernet, terrestrial/satellite Four HDMI 2.1, digital audio output, two USB ports, Ethernet, terrestrial/satellite Four HDMI, digital audio output, two USB ports, Ethernet, Terrestrial/satellite Four HDMIs (two with full HDMI 2.1 features), USB 3.0, Ethernet, RF input, optical digital audio output Four HDMI, digital audio out, two USB ports, Ethernet, two satellite, RF terrestrial HDMI (2.1) eARC, ALLM, VRR, HFR eARC, ALLM, HRF, VRR VRR, eARC, ALLM, 4K/120Hz Four ports with 4K/120, ALLM and VRR support eARC, ALLM, VRR, HFR eARC, ALLM, VRR, HFR, QMS eARC, VRR, ALLM, 4K/120Hz eARC, ALLM, VRR, HFR eARC, ALLM, VRR, HFR eARC, ALLM, VRR, HFR eARC, ALLM, VRR, HFR, SBTM Audio (Power output) 40 W 70 W 50 W 40 W 40 W 20 W 60 W 20 W 40 W 60 W 50 W Connectivity Wi-Fi, Bluetooth 5.1, AirPlay 2 Wi-Fi, Bluetooth 5.3, Apple Airplay 2 Wi-Fi, Bluetooth, Chromecast, AirPlay 2 Three USBs, Four HDMIs, optical digital audio output, Ethernet port, RF port, Wi-Fi, Bluetooth Wi-Fi, AirPlay 2, Bluetooth Wi-Fi, Bluetooth 5.3, AirPlay 2, Google Cast Wi-Fi, Bluetooth Wi-Fi, Bluetooth – Wi-Fi, Bluetooth 5.4, Miracast Wi-Fi, AirPlay 2, Google Cast, Bluetooth 5.3 Colours – Black – Silver – Black black – Black – – Display Technology OLED Mini LED OLED Direct-LED (Full Array Local Dimming) Mini LED OLED OLED OLED OLED, QLED Mini LED OLED

Anthropic on Monday launched the most ambitious consumer AI agent to date, giving its Claude chatbot the ability to directly control a user’s Mac — clicking buttons, opening applications, typing into fields, and navigating software on the user’s behalf while they step away from their desk.

The update, available immediately as a research preview for paying subscribers, transforms Claude from a conversational assistant into something closer to a remote digital operator. It arrives inside both Claude Cowork, the company’s agentic productivity tool, and Claude Code, its developer-focused command-line agent. Anthropic is also extending Dispatch — a feature introduced last week that lets users assign Claude tasks from a mobile phone — into Claude Code for the first time, creating an end-to-end pipeline where a user can issue instructions from anywhere and return to a finished deliverable.

The move thrusts Anthropic into the center of the most heated competition in artificial intelligence: the scramble to build agents that can act, not just talk. OpenAI, Google, Nvidia, and a growing swarm of startups are all chasing the same prize — an AI that operates inside your existing tools rather than beside them. And the stakes are no longer theoretical. Reuters reported Sunday that OpenAI is actively courting private equity firms in what it described as an “enterprise turf war with Anthropic,” a battle in which the ability to ship working agents is fast becoming the decisive weapon.

The new features are available to Claude Pro subscribers (starting at $17 per month) and Max subscribers ($100 or $200 per month), but only on macOS for now.

Inside Claude’s computer use: How Anthropic’s AI agent decides when to click, type, and navigate your Mac

The computer use feature works through a layered priority system that reveals how Anthropic is thinking about reliability versus reach.

When a user assigns Claude a task, it first checks whether a direct connector exists — integrations with services like Gmail, Google Drive, Slack, or Google Calendar. These connectors are the fastest and most reliable path to completing a task, according to Anthropic’s documentation. If no connector is available, Claude falls back to navigating the Chrome browser via Anthropic’s Claude for Chrome extension. Only as a last resort does Claude interact directly with the user’s screen — clicking, typing, scrolling, and opening applications the way a human operator would.

This hierarchy matters. As Anthropic’s help center documentation explains, “pulling messages through your Slack connection takes seconds, but navigating Slack through your screen takes much longer and is more error-prone.” Screen-level interaction is the most flexible mode — it can theoretically work with any application — but it is also the slowest and most fragile.

When Claude does interact with the screen, it takes screenshots of the user’s desktop to understand what it’s looking at and determine how to navigate. That means Claude can see anything visible on the screen, including personal data, sensitive documents, or private information. Anthropic trains Claude to avoid engaging in stock trading, inputting sensitive data, or gathering facial images, but the company is candid that “these guardrails are part of how Claude is trained and instructed, but they aren’t absolute.”

There is nothing to configure. No API keys, no terminal setup, no special permissions beyond what the user grants on a per-app basis. As Ryan Donegan, who handles communications for Anthropic, put it in a press briefing: “Download the app and it uses what’s already on your machine.”

Claude Dispatch turns your iPhone into a remote control for AI-powered desktop automation

The real strategic play may not be computer use itself but how Anthropic is pairing it with Dispatch.

Dispatch, which launched last week for Cowork and now extends to Claude Code, creates a persistent, continuous conversation between Claude on your phone and Claude on your desktop. A user pairs their mobile device with their Mac by scanning a QR code, and from that point forward, they can text Claude instructions from anywhere. Claude executes those instructions on the desktop — which must remain awake and running the Claude app — and sends back the results.

The use cases Anthropic envisions range from mundane to ambitious: having Claude check your email every morning, pull weekly metrics into a report template, organize a cluttered Downloads folder, or even compile a competitive analysis from local files and connected tools into a formatted document. Scheduled tasks allow users to set a cadence once — “every Friday,” “every morning” — and let Claude handle the rest without further prompting.

Anthropic’s blog post frames the combination of Dispatch and computer use as something of a paradigm shift. “Claude can use your computer on your behalf while you’re away,” the company wrote, offering examples like creating a morning briefing while a user commutes, making changes in an IDE, running tests, and submitting a pull request.

One early user on social media captured the broader ambition succinctly. Gagan Saluja, who describes himself as working with Claude and AWS, wrote: “combine this with /schedule that just dropped and you’ve basically got a background worker that can interact with any app on a cron job. that’s not an AI assistant anymore, that’s infrastructure.”

First hands-on tests reveal Claude’s computer use works about half the time — and that may be the point

Anthropic is calling this a research preview for a reason. Early hands-on testing suggests the feature works well for information retrieval and summarization but struggles with more complex, multi-step workflows — particularly those that require interacting with multiple applications.

John Voorhees of MacStories, the Apple-focused publication, published a detailed hands-on evaluation of Dispatch the same day as the announcement. His results were mixed. Claude successfully located a specific screenshot on his Mac, summarized the most recent note in his Notion database, listed notes saved that day, added a URL to Notion, summarized his most recently received email, and recalled a screenshot from earlier in the session. But it failed to open the Shortcuts app on his Mac, send a screenshot via iMessage, list unfinished Todoist tasks (due to an authorization error), list Terminal sessions, display a food order from an active Safari tab, or fetch a URL from Safari using AppleScript.

Voorhees’ verdict was measured: Dispatch “can find information on your Mac and works with Connectors, but it’s slow and about a 50/50 shot whether what you try will work.” He added that it is “not good enough to rely on when you’re away from your desk” but called it “a step in the right direction.”

Meanwhile, on GitHub, users are already surfacing technical issues. One bug report filed against Claude Code describes a scenario where the Read tool attempts to process multiple large PDF files in a single turn without checking whether the combined payload exceeds the 20MB API limit, causing the request to fail outright. The issue, which has been tagged as a bug specific to macOS, highlights the kinds of rough edges that come with shipping an early preview of a complex agentic system.

OpenClaw, NemoClaw, and the startup swarm: Why Anthropic is racing to ship AI computer use now

Anthropic’s timing is not accidental. The company is shipping computer use capabilities into a market that has been rapidly reshaped by the viral rise of OpenClaw, the open-source framework that enables AI models to autonomously control computers and interact with tools.

OpenClaw exploded earlier this year and proved that users wanted AI agents capable of taking real actions on their computers — and that they were willing to tolerate rough edges to get them. The framework spawned an entire ecosystem of derivative tools — what the community calls “claws” — that turned autonomous computer control from a research curiosity into a product category almost overnight. Nvidia entered the fray last week with NemoClaw, its own framework designed to simplify the setup and deployment of OpenClaw with added security controls. Anthropic is now entering a market that the open-source community essentially created, betting that its advantages — tighter integration, a consumer-friendly interface, and an existing subscriber base — can compete with free.

Smaller startups are also pushing into the space. Coasty, which offers both a desktop app and browser-based AI agent for Mac and Windows, markets itself as providing “full browser, desktop, and terminal automation with a native experience.” One user on social media directly pitched Coasty in the replies to Anthropic’s announcement, claiming it offers “much better user experience and more accurate” results — a sign of how crowded and competitive the computer-use agent space has become in a matter of months.

The competitive dynamics extend beyond just computer use. Reuters has reported that OpenAI is sweetening its pitch to private equity firms amid what the wire service described as an “enterprise turf war with Anthropic.” The two companies are locked in an escalating battle for enterprise customers, and the ability to offer agents that can actually operate within a company’s existing software stack — not just chat about it — is increasingly the differentiator.

Prompt injection, screenshot surveillance, and the unsolved security risks of letting AI control your desktop

If the competitive pressure explains why Anthropic shipped this feature now, the safety caveats explain why the company is hedging its bets.

Computer use runs outside the virtual machine that Cowork normally uses for file operations and commands. That means Claude is interacting with the user’s actual desktop and applications — not an isolated sandbox. The implications are significant: a misclick, a misunderstood instruction, or a prompt injection attack could have real consequences on a user’s live system.

Anthropic has built several layers of defense. Claude requests permission before accessing each application. Some sensitive apps — investment platforms, cryptocurrency tools — are blocked by default. Users can maintain a blocklist of applications Claude is never allowed to touch. The system scans for signs of prompt injection during computer use sessions. And users can stop Claude at any point.

But the company is remarkably forthright about the limits of these protections. “Computer use is still early compared to Claude’s ability to code or interact with text,” Anthropic’s blog post states. “Claude can make mistakes, and while we continue to improve our safeguards, threats are constantly evolving.”

The help center documentation goes further, explicitly warning users not to use computer use to manage financial accounts, handle legal documents, process medical information, or interact with apps containing other people’s personal information. Anthropic also advises against using Cowork for HIPAA, FedRAMP, or FSI-regulated workloads.

For enterprise and team customers, there is an additional wrinkle. Cowork conversation history is stored locally on the user’s device, not on Anthropic’s servers. But critically, enterprise features like audit logs, compliance APIs, and data exports do not currently capture Cowork activity. This means that organizations subject to regulatory oversight have no centralized record of what Claude did on a user’s machine — a gap that could be a dealbreaker for compliance-sensitive industries.

One user flagged this concern on social media with particular precision. NomanInnov8 wrote: “when the agent IS the user (same mouse, keyboard, screen), traditional forensic markers won’t distinguish human vs AI actions. How are we thinking about audit trails here?”

The question is not academic. As AI agents gain the ability to take real-world actions — sending emails, modifying files, interacting with financial systems — the ability to distinguish between human and machine actions becomes a foundational requirement for governance, liability, and compliance. Anthropic has not yet answered it.

From excitement to anxiety: How users are reacting to Claude’s new power over their machines

The social media reaction to the announcement split roughly into three camps: those excited about the productivity implications, those concerned about the security risks, and those frustrated that they cannot yet use it.

The enthusiasm was genuine and widespread. “Legit just got the update and used it with dispatch — exactly the feature I wanted,” wrote one X user. Mike Joseph called the speed of Anthropic’s feature releases “fantastic.” Another X user noted the significance for non-technical users: “Very exciting for non-tech folks who don’t want or know how to set up OpenClaw.”

But the security concerns were equally pointed. One user, posting as Profannyti, wrote: “Granting that kind of control over your personal device doesn’t sit right. It’s almost like letting someone you barely know take the wheel and trusting everything will be fine.” 

As Engadget reported, experts have warned that one major concern with agentic AI is that “it can take major, sometimes dramatic actions quickly and with little warning,” and that such tools “can also be hijacked by malicious actors.”

Several users flagged practical frustrations as well. Windows users — excluded from the macOS-only research preview — expressed predictable dismay. Others reported that the new features were consuming their usage quotas at alarming rates. One Max 20x subscriber paying $200 per month complained that Dispatch was “eating my quota like crazy,” consuming 10% of their allowance in a single prompt. Another user linked to the GitHub bug report about the 20MB payload issue, calling the situation “quite urgent.”

Anthropic’s enterprise playbook: Plugins, pricing tiers, and the bet that AI agents can replace entire workflows

The pricing structure reveals where Anthropic sees the real market. While individual Pro users get access to Cowork, the company notes that agentic tasks “consume more capacity than regular chat” because “Claude coordinates multiple sub-agents and tool calls to complete complex work.” Heavy users are nudged toward Max plans at $100 or $200 per month.

For teams, the pricing starts at $20 per seat per month for groups of five to 75 users. Enterprise pricing is custom and includes admin controls to toggle Cowork on or off for the organization.

The plugin architecture is where Anthropic’s enterprise ambitions become clearest. Plugins bundle skills, connectors, and sub-agents into a single install that turns Claude into a domain specialist — for legal work, finance, brand voice management, or other functions. Anthropic already lists plugins for legal workflows (contract review, NDA triage), finance (journal entries, reconciliation, variance analysis), and brand voice (analyzing existing documents to enforce guidelines). The company is betting that the combination of computer use, Dispatch, scheduled tasks, and domain-specific plugins will create an agent capable enough to justify enterprise procurement.

The testimonials Anthropic has gathered suggest the pitch is landing with at least some organizations. Larisa Cavallaro, identified as an AI Automation Engineer, described connecting Cowork to her company’s tech stack and asking it to identify engineering bottlenecks. Claude, she said, returned “an interactive dashboard, team-by-team efficiency analyses, and a prioritized roadmap.” Joel Hron, a CTO, offered a more philosophical framing: “The human role becomes validation, refinement, and decision-making. Not repetitive rework.”

The AI industry’s defining tension: Shipping fast enough to win, slow enough to be safe

Anthropic is shipping these capabilities at a moment of extraordinary velocity in the AI industry — and extraordinary uncertainty about what that velocity means.

The company’s own research quantifies the transformation underway. Its economic index, published in March 2026, tracks how AI is reshaping labor markets and productivity across sectors. The data suggests that AI adoption is accelerating unevenly, with knowledge workers in technology, finance, and professional services seeing the most dramatic shifts.

Anthropic is also navigating significant external pressures beyond the product arena. Recent reporting has highlighted scrutiny from Senator Elizabeth Warren regarding Anthropic’s defense and supply chain relationships — a reminder that the company’s ambitions to build powerful autonomous agents exist within an increasingly complex political and regulatory environment.

For now, the computer use feature remains early and imperfect. Complex tasks sometimes require a second attempt. Screen interaction is meaningfully slower than direct integrations. The audit trail gap for enterprise users is a genuine liability. And the fundamental tension between giving an AI agent enough access to be useful and limiting that access enough to be safe remains unresolved.

But Anthropic is not waiting for perfection. The company is building in public, shipping capabilities it openly describes as incomplete, and betting that users will tolerate a 50 percent success rate today in exchange for the promise of something transformative tomorrow. It is a calculation that only works if the failures remain minor — a missed click, a stalled task, an unread email. The moment a failure isn’t minor, the calculus changes entirely.

The AI industry has spent the last three years proving that machines can think. Anthropic is now asking a harder question: whether humans are ready to let them act. The answer, for the moment, is a provisional yes — hedged with permissions dialogs, blocklists, and the quiet hope that nothing important gets deleted before the technology catches up to the ambition.