In addition to the company's GPT-6.1 Astra announcement, OpenAI announced always-on AI agents called "Dots" that can proactively use a computer and connected apps to research, draft documents, write software and handle other tasks on a user's behalf. "OpenAI said it would roll out Dots to paid users starting Tuesday," reports Bloomberg. "Initially, OpenAI said, each user will have access to a single 'dot' that can act as an assistant. Over time, the company expects users to manage a slew of them." From the announcement: Dots are frontier intelligence that have your back. Powered by GPT-6 Astra, they have their own cloud computer, learn from feedback over time, and can work towards your goals 24/7. Through our ecosystem of plugins, they can readily connect to over 4,000 apps, giving them the tools to help wherever you need them.

Your dot is always within reach, whether it is through ChatGPT, Slack, or Teams. You can ask questions, explore ideas, and give feedback along the way; or you can simply hop on a voice call when you need to talk something out. The more you work together, the more your dot learns your preferences, how you think, and what good looks like to you.

The magic of dots is when they bring you work done the way you would do it, sometimes before you even think to ask. With dots alongside you, ambitious ideas and projects you haven't had the time or support to pursue start to feel achievable.

The United States and China have agreed to "launch a dialogue" on AI, reports Reuters. On artificial intelligence, the two sides agreed to hold a dialogue on the technology's risks and benefits, with the next round of discussions set for November, and to set up a communication channel for AI-related incidents, the Chinese Foreign Ministry and the White House said.

The White House said that the leaders had agreed to use the term "super intelligence" in place of "artificial intelligence." In a separate statement, the Chinese ministry said that Beijing valued Washington's use of the new term. As AI technology continues to advance, the two sides should step up exchanges and work toward consensus in line with new developments, it said.
But CNN argues that "Despite growing calls to prevent AI development from spiraling out of control, the Trump-Xi summit has produced little substance, as many experts expected." The right thing to do on AI, [China's leader] Xi said during talks with Trump, is to "draw on each other's strengths, not guard against each other" — a reference to Beijing's concern about US containment, from existing tech export controls to potential AI restrictions. "The two sides can continue their dialogue on AI, exchange views on its risks and benefits, and jointly prevent the misuse and abuse of AI," he added. But the summit has yielded little progress on AI beyond a formal dialogue and a bilateral communication channel, proposals discussed before the two leaders' summit — underscoring the entrenched mutual mistrust amid contrasting visions on AI... Because of low levels of trust, cooperation between the two superpowers remains limited, said George Chen, chair of digital practice at The Asia Group consultancy. "Beijing continues to believe Washington seeks to contain China's rise in AI and other emerging technologies, a perception that will shape the pace and scope of future engagement for the two countries on AI," he said.
CNN also points out that while China trails the US in frontier AI models, "it's rapidly narrowing the technology gap while championing a more open ecosystem centered on accessibility and lower cost." In July, Chinese leader Xi Jinping launched the World Artificial Intelligence Cooperation Organization — a rival grouping to the Pax Silica alliance that Trump formed last year to reduce reliance on China for AI supply chains. While over two dozen countries and the European Union signed up to Trump's Pax Silica, Xi has recruited 29 countries, including Russia, Indonesia and Pakistan, to his alternative vision of open models, which allow users to freely download, customize and run without paying hefty fees to American firms like Anthropic and OpenAI. For developers in the Global South, an inexpensive Chinese model from DeepSeek or Moonshot may be more useful than a slightly more capable system requiring an expensive subscription and access to a foreign cloud provider, said Eric Olander, editor in chief of The China-Global South Project, a research agency....

China's embrace of open systems has not always been a top-down strategy by Beijing. Restrictions on access to the most advanced chips because of US export controls, coupled with smaller capital markets, have pushed Chinese developers toward open models as a way to compete with leading US proprietary systems. That shift has proved effective. In a year, Chinese models' global usage skyrocketed from less than 15% to over 54% last week, led by DeepSeek, according to AI leaderboard data by OpenRouter, a marketplace for models. Even American firms, from Airbnb and DoorDash to Shopify, have embraced Chinese models, tapping into the advantages of open systems, including lower costs and greater flexibility for customization.
CNN adds this insight from Alex Colville, an analyst focusing on tech and security at the government-backed Australian Strategic Policy Institute. "The more capable Chinese models become, the less likely it is Beijing may leave them unrestricted."

Microsoft's security blog describes the fight against a new "AI-powered cybercrime platform" offering phishing-as-a-service, with AI-tailored lures and analyses of compromised inboxes (to identify high-value targets). The site compromised more than 12,000 inboxes in over 10,000 organizations around the world, compromising business accounts "at scale" with automated attacks and prebuilt phishing templates. AI tools could even sift through a victim's mailbox to help engineer better phishing messages.

To disrupt EvilTokens Microsoft worked with other organizations, including Cloudflare, Coinbase, OpenAI, Railway, SpyCloud, Shadowserver Foundation, and TRM Labs to Health-ISAC (a non-profit helping health sector organizations share cyber threat information).

"Fifty sites seized and 150 domains disabled in a single action is only possible when the hosting providers, the exchanges, the model providers and the data holders all move at the same time," security company SpyCloud told The Hacker News.

From Microsoft's security blog: Microsoft also notified affected customers, helped remediate compromised accounts, and shared intelligence to support further defensive and investigative action... Microsoft worked closely with specialist officers from the Metropolitan Police Service's cybercrime team, sharing intelligence that enabled officers to take operational action in the United Kingdom. On September 11, 2026, officers arrested two men, aged 32 and 38, and seized digital devices and other items for examination... While EvilTokens used AI to identify targets and prioritize fraud opportunities, Microsoft investigators used reverse engineering and AI-powered tools to analyze evidence, accelerate the investigation, and identify the infrastructure supporting the service...

Campaigns leveraging EvilTokens have impacted organizations in various industries, including wholesale distribution, construction, financial services, real estate, higher education, and healthcare, with the highest concentrations of observed victim activity in the United States, Canada, the United Kingdom, Australia, India, and France. Working with partners, Microsoft's Digital Crimes Unit (DCU) facilitated a coordinated disruption of infrastructure used to operate the EvilTokens service.
Sometimes stolen tokens were used to give new devices access to a victim's inbox. (A code authenticating the new device was sent to the targeted user, who unknowingly authorize the threat actor's session and grants access to their account...) But "AI was not simply helping attackers write more convincing messages," says another Microsoft blog post. "It helped them decide who to target, who to impersonate, and how to most effectively exploit the relationship to extract as much money as possible." The significance of EvilTokens extends beyond its rapid growth and global reach. It offers an early warning of what happens when cybercriminals combine stolen access with AI capable of understanding how an organization works... Its AI tools could summarize and translate emails, surface financial conversations, map organizational roles, identify trusted relationships, and recommend potential targets. Preset prompts offered to find wire-transfer discussions, identify the organization's "money movers," locate vendor invoices, and determine the best people to impersonate. Sold through Telegram for a $1,500 initiation fee and a recurring $500 subscription, EvilTokens combined account compromise, mailbox analysis, target selection, and fraud preparation in a single service. Capabilities that once required experience across identity attacks, cloud systems, social engineering, and financial fraud were available through a ready-made interface.

Investigators found evidence that large portions of EvilTokens had been "vibe coded," with AI helping its creators build the platform itself. They also determined that EvilTokens drew on capabilities from multiple AI models. The result was more than a collection of attack tools. EvilTokens packaged much of the fraud process into a commercially run service, complete with subscription pricing, customer support, management dashboards, and tools designed to move customers from account access toward financial exploitation.

Ars Technica reports: Half a year after Iranian drone strikes knocked out multiple Amazon data centers, the US tech company has acknowledged the permanent loss of some customer data that was hosted in Bahrain and the United Arab Emirates.

Amazon Web Services was "unable to restore access to the resources and data" hosted in some of the war-damaged data centers, according to an AWS dashboard update posted on September 15. The development was first reported by Reuters and suggests that the Iranian strikes inflicted catastrophic damage on the data centers. Customer data was irretrievably lost in one of three AWS availability zones in the United Arab Emirates region, specifically the mec1-az2 availability zone.

Each availability zone is serviced by one or more Amazon data centers. The company is still working on "recovering regional resources" and restoring resources hosted in the other two availability zones in the United Arab Emirates... But the destruction was apparently even more widespread for the data centers in the Bahrain region, because Amazon said it was unable to restore access to resources and data across all three availability zones there. "The damage to our infrastructure spanned multiple Availability Zones and exceeded what our regional and multi-AZ services are designed to withstand," the AWS update said.

The company had already suspended customer billing in the affected AWS regions in Bahrain and the United Arab Emirates and has spent the past six months attempting to restore normal operations... [When the war began] AWS immediately urged customers to migrate resources to other cloud regions and use remote backups to restore any "inaccessible resources." The cloud service also reportedly issued $150 million in customer credits following the initial Iranian strikes in March.

The NSA, CISA, and FBI are accusing (PDF) several Chinese AI companies of carrying out "industrial-scale" distillation campaigns against leading U.S. models such as ChatGPT, Claude, Gemini, and Grok. Since at least 2024, the companies have allegedly routed millions of requests across accounts, APIs, proxies, cloud providers, and third-party aggregators to extract capabilities for their own models. "China-based artificial intelligence companies are conducting systematic extraction of proprietary functionalities and capabilities of U.S. AI companies' models through industrial-scale knowledge distillation campaigns that form the core -- not merely a supplement -- of their AI development strategy," the agencies wrote. CyberScoop reports: DeepSeek, for example, distilled frontier U.S. models to generate synthetic training data for its R1 and R3 models, including four different versions of Claude, two versions of Gemini, five versions of ChatGPT and Grok 4. Those models helped train DeepSeek's capabilities in areas like agentic functioning, question and answer optimization, creative and occupational writing and others.

Another Chinese company, Moonshot AI, allegedly distilled 18 different U.S. models -- including Fable 5, Anthropic's current, most advanced commercially available model -- to train its Kimi-K2 and Kimi K3 models. The company used millions of queries meant to extract enhanced capabilities in areas like agentic reasoning, coding and data analysis, computer vision, larger logical frameworks, visual processing and others.

Chinese AI companies manage a sophisticated set of tools and systems that route requests and prompts through multiple pathways to avoid detection. The advisory lists common tactics observed by Chinese companies, including spreading requests across different accounts, models and platforms, using native APIs, remote cloud providers, and third-party aggregators to obfuscate user metadata, and leveraging proxies and gray tech markets to get around geographic restrictions, terms of use and safeguards built into frontier models.

Meta has launched Muse, a personal AI agent developed under chief AI officer Alexandr Wang. "The product, long in development, was touted as a key next step by CEO Mark Zuckerberg in his recent 6,500-word manifesto," reports Axios. From the report: Muse, as the agent is known, exists in a chat interface, similar to a text thread. It's designed to be more proactive and long-running than typical chatbots. Users can name their agent, create an avatar and customize how it communicates. The Muse agent runs on a dedicated virtual machine in Meta's cloud, using a built-in browser that's visible to the user.

Meta is offering a free tier of Muse, as well as two subscription options, at $20 per month and $100 per month. "For the vast majority of users, they should be able to do what they need to within the free tier," Wang told Axios. "But for real power users, you know, those subscription tiers help us cover the computer costs." There is no advertising within Muse, but Wang said the company is exploring commerce opportunities that could generate additional revenue. Initially Muse will be available in the U.S. and works on iOS, Android and the web, with support coming soon for Meta's AI glasses. "The full vision in the future is we want to develop personal superintelligence that helps people accomplish their goals, pursue their passions, build things that they never would have built if they didn't have the technology," Wang told Axios.

Meta offers users more privacy controls with Muse than in its previous AI products, including the option to prevent queries from being used by Meta and a planned confidential mode where the company cannot see activity inside a user's virtual workspace. There's also an entirely separate system called Sentinel that governs Muse's access to the internet and connected services.

Google DeepMind has released AlphaGenome Atlas, a free research database containing AI-generated predictions for the effects of all 9 billion possible single-letter mutations in the human genome. Built from its AlphaGenome model, the atlas is designed to help scientists interpret both protein-coding and harder-to-understand regulatory DNA. Fortune reports: AlphaGenome Atlas, as DeepMind calls the database, is a precomputed catalogue of what each substitution of a single DNA base is likely to do to the machinery that switches genes on and off. Until now researchers had to run such a model one variant at a time or had to test variants in the laboratory, a process that was painstakingly slow. It would have taken many human lifetimes to discover the consequences of all 9 billion possible single-letter mutations. The Atlas promises to make the job of biologists and medical researchers considerably easier, potentially speeding up the understanding of genetic diseases and the hunt for possible cures.

Pushmeet Kohli, DeepMind's vice president for research and head of its AI for science team, told reporters on a briefing call that this was the first time any researcher in the world could reach a comprehensive map of human genetic variation "by simply opening a browser." Kohli also framed the release as helping to complete the unfinished business of the Human Genome Project, which in 2003 succeeded in mapping the entire human DNA sequence. "As the saying goes, we bought the book," he said, "but we did not understand how to read it."

Atlas is available for non-commercial use from today through a website Google DeepMind has set up for it. The company said it would be available for commercial use through a licensing arrangement through Google Cloud "soon." Kohli said that Google DeepMind's sister company,ÂIsomorphic Labs, which is using AI for drug discovery, would have access to Atlas but that it would also require a commercial license for access. He did not specify exactly what the terms would be for commercial licensing. A paper describing the Atlas and how it was created is being released on bioRxiv, a repository for biomedical preprint academic papers.

Switzerland has a population of 9,154,242. And its federal government just launched a pilot program to test open source alternatives to Microsoft 365 on 3,000 workstations, reports the blog It's FOSS. "That's about 7% of the federal workforce." The target is to complete the migration by end of 2027... On September 3, 2026, the Federal Council published results of "PoC BOSS", a feasibility proof-of-concept involving 172 federal employees who tested the openDesk suite, a German open-source collaboration platform. During the proof-of-concept phase, core office tasks like document processing and email received positive assessments, while large-scale video conferencing still showed technical limitations. Based on the 'success' of the PoC phase with 172 employees, the pilot is now launched for 3,000 employees...

According to Matthias Stürmer, professor at the Bern University of Applied Sciences, Microsoft's supremacy in public institutions poses three problems that are driving this migration. First is the risk of foreign access. US cloud legislation could expose Swiss government data to foreign authorities. Second is the risk to service continuity, as dependency on a single foreign vendor creates operational risk. The third risk is the escalating costs as proprietary licensing fees are rising with no Swiss leverage.

Switzerland's military cybersecurity unit, Cyber Command, is not waiting for the civilian pilot. It is already poised to replace Microsoft 365 entirely with openDesk by October 2026. It is pretty much the same reason. Military doesn't want foreign governments accessing sensitive Swiss data.
The article notes that during the pilot phase, the new system runs in parallel with Microsoft 365 rather than replacing it. But the article's author speculates that "If the pilot is successful, we might expect the migration to continue on all the 54,000 workstations owned by the federal administrations."

Microsoft is imposing monthly limits on Xbox Cloud Gaming starting in November, with Game Pass Ultimate, Premium, and Essential subscribers getting 15, 10, and 5 hours respectively before they have to buy additional playtime. "Currently, Game Pass subscribers can stream without these monthly hour limits," notes IGN. From the report: The change means those who use cloud gaming most via Game Pass will end up having to pay more, although Microsoft said this impacts 4% of Game Pass subscribers. Also in November, Microsoft said players will be able to buy cloud playtime hours through the Xbox Store and 'stream eligible games they own on supported devices' without subscribing to Game Pass.

In a blog post, Microsoft said it was making the changes because the cost of providing cloud gaming has grown as more people use it and play for longer. The monthly limits, Microsoft continued, will mean it can invest in reliability and performance. But it said it understood that for some players the practical result is a higher cost.

Dyson is entering oral care with the $499 CameraJet, a smart toothbrush that uses a tiny camera and on-device AI to detect gaps between teeth and automatically blast them with mouthwash while you brush. Regular toothpaste and mouthwash will work with the CameraJet, but Dyson plans to sell its own non- or low-foaming formulas designed to give the toothbrush's camera a clearer view inside your mouth. The Verge reports: Just below the brush head is a tiny camera with a 100,000-pixel resolution and a strobing light. The toothbrush's AI analyzes 28 live images from the camera every second as part of a feature Dyson calls Gap Optical Targeting, without storing them on the device or sending them to the cloud. Within 100 milliseconds of a gap between your teeth being detected, the CameraJet emits a cone-shaped blast of mouthwash to clean out the space. Dyson uses a wider blast than what water flossers typically employ so it can effectively remove plaque with less pressure, making the CameraJet's flossing mechanism gentler on enamel and gums

The replaceable brush heads use a combination of softer outer bristles that safely clean along the gumline and stiffer inner bristles that are better at scrubbing away tooth stains. The brush head oscillates nearly 1,000 times each second while the angle of oscillation is constantly changing to improve the effectiveness of your brushing. Pricing on replacement brush heads isn't known, but each one features an RFID tag so the CameraJet can track usage and alert you when a replacement is needed every three months.

To ensure its mouthwash reservoir can effectively pump no matter how the CameraJet is held -- even upside down -- instead of sticking a tube into the reservoir, there's an air bladder inside the tank that slowly inflates to fill empty space, minimize air pockets, and maintain a constant pressure. An elaborate docking station recharges the toothbrush and can automatically fill its built-in 12.5ml mouthwash reservoir in just three seconds.

An anonymous reader quotes a report from VentureBeat: The viral fervor we saw earlier this year around OpenClaw, the open source AI harness that turns powerful language models into autonomous workers the user can message via their favorite channels (Telegram, iMessage, WhatsApp, Discord etc), has cooled off substantially from its peak in March 2026. But over the weekend, OpenClaw's creator Peter Steinberger and current team of co-developers gave the world -- especially enterprises -- a reason to look at it again, announcing OpenClaw 2.0, billed as the most significant update to the harness and surrounding platform yet.

OpenClaw 2.0 seeks to transform what began largely as a personal agent harness into something increasingly designed for teams, shared infrastructure and enterprise workflows. OpenClaw 2.0 introduces a rebuilt browser interface that brings conversations, files, approvals, configuration and live agent activity into a common workspace. It adds shared cloud sessions and multi-user collaboration. And it expands the security model with stronger sandboxing, role-based permissions, approval controls, secrets handling and auditing.

Together, those additions move OpenClaw closer to being infrastructure that an organization could deploy for employees rather than simply a powerful agent an individual developer runs locally. They also sharpen a competitive question surrounding the project: whether OpenClaw has addressed the security and isolation concerns that helped inspire newer alternatives such as NanoClaw. The answer is increasingly yes at the capability level -- but not necessarily by default. Over the past two months, OpenClaw has been "build[ing] OpenClaw with OpenClaw," shifting its team from individual local coding tools to a shared agent environment at team.openclaw.ai that can see what everyone is working on. Steinberger called "multiplayer coding + infinite compute with nodes and cloud sessions" a "game changer," adding that local coding harnesses now "feel like relics of the past."

VentureBeat highlights a use case from Solvely CEO Colin Johnson that shows how OpenClaw's multiplayer WebUI can improve collaboration between developers and AI agents. Instead of simply "messaging a bot," multiple developers can now join the same live session, see the same history and artifacts, and work "inside the same context." In Johnson's case, that meant a developer taking over his project could simply join the existing agent thread rather than rely on a separate handoff document. As he put it, "the session itself became the handoff document."

Anthropic has reportedly signed a $35 billion cloud-computing deal with Nvidia-backed Lambda for a roughly 350-megawatt Texas data center being developed by Hut 8. "Lambda will install Nvidia chips at the Hut 8 facility, with the arrangement designed to expand the pool of Nvidia computing resources available to power Anthropic's Claude AI products," reports Quartz. From the report: The arrangement is a further sign of how Nvidia is positioning itself as a facilitator of computing access across the AI industry. It also benefited Lambda, which counts Nvidia as an investor and was able to land a major Anthropic contract without having to independently source or lease the underlying data center space, according to the Journal.

The data center covers roughly 350 megawatts of capacity, according to Bloomberg. Hut 8, which began as a bitcoin mining company before moving into AI data centers, disclosed in July that an investment-grade customer had signed 15-year leases for its Texas campus, with a base contract value of $19.6 billion, without naming the tenant.

The Lambda deal is the latest in a series of large computing agreements Anthropic has pursued. Last week, the company agreed to spend $45 billion to rent capacity from Nscale's West Virginia data center campus. It has also signed deals worth $50 billion with neocloud Fluidstack Ltd. and $45 billion with SpaceX, according to Bloomberg.

TIME's 2026 list of the 100 most influential people in AI omits Nvidia CEO Jensen Huang and AMD CEO Lisa Su, despite both having appeared on previous editions and playing central roles in the hardware powering the AI boom. Nvidia's only representative this year is sustainability chief Josh Parker, while the list includes figures such as Ben Affleck and Paris Hilton alongside major tech executives and policymakers. Tom's Hardware reports: Broadcom CEO Hock Tan, Micron CEO Sanjay Mehrotra, SMIC co-CEO Liang Mong Song, and Huawei HiSilicon president He Tingbo all appear on the 2026 list, joined by Cerebras' Andrew Feldman, CoreWeave's Michael Intrator, and Nscale's Josh Payne. Together, the roster includes Nvidia's HBM supplier, two of its biggest GPU cloud customers, its custom-accelerator rivals, and China, which is trying -- and so far struggling -- to replace Nvidia altogether.

Parker's profile, which appears under the Thinkers category, focuses on data center energy and water consumption, alongside Nvidia's tripling of Scope 3 emissions in two years, and the Rubin platform's shift to 100% liquid cooling. "Nvidia is at the center of the AI revolution," Parker told TIME.

TIME editor-in-chief Sam Jacobs wrote in the list's accompanying article that selection was led by editor Ayesha Javed, and that the 100 honorees represent the year's "key storylines and who, in our opinion, are having the most influence in driving these developments." Now in its fourth year, this edition of the TIME100 AI appears to have focused especially on featuring the new additions to TIME's community of AI leaders, with Mark Zuckerberg, Demis Hassabis, Sundar Pichai, and Satya Nadella representing four other honorees, arguably deserving of a space on this year's list, churned out.

Su's strong TIME back catalog makes her omission all the stranger still. The magazine named her its 2024 CEO of the Year, put her on the 2024 AI list and the main TIME100 in 2025, and included her among December's Architects of AI.

According to The Information (paywalled), Apple reportedly moved up the release of new Mac mini and Mac Studio models after unexpectedly strong enterprise demand for Macs capable of running AI workloads. MacRumors reports: Apple normally releases new Mac models in the autumn, closer to October or November, making this week's announcement unusually early, falling just before the anticipated arrival of new iPhone models. The Information says that the AI-driven boom in Mac Studio and Mac mini sales is behind the early launch. Apple noticeably promoted the ability to link multiple Mac Studios together into a single, more capable system for running large frontier AI models, a feature aimed at business and developer customers rather than everyday consumers.

Apple highlighted the Mac mini and Mac Studio's shift toward business buyers in June, with a "Business at the Park" event involving executives from major companies Ford, Disney, and Anthropic. The Mac mini was said to be the "darling" of the event. Even so, enterprise's rush toward powerful desktop Macs more broadly took Apple by surprise.

The company reportedly did not possess an engineering team dedicated to business customers or staff focused on developer relations, and lacked an enterprise AI strategy. Businesses that approached Apple asking to buy access to the company's Private Cloud Compute infrastructure were reportedly turned down. Apple is instead leaning on partners such as WebAI and Mount Thor, which provide AI tools and execution environments built on Apple hardware.

Citrix's Desktop as a Service (DaaS) product includes a lightweight, hardened second operating system called UniconOS (once called "eLux") to offer Windows customers a write-protected file system Citrix says is secure against computer viruses and other malware. Nerds.xyz reports: According to the company, the environment remains isolated from the Windows filesystem and uses Secure Boot protections covering the shim, bootloader, kernel, and initial RAM filesystem. That separation matters when ransomware encrypts Windows or a faulty update leaves the operating system trapped in a boot loop. Instead of repairing Windows immediately, an employee can boot into UniconOS and use Citrix DaaS to access virtual applications and desktops.

Citrix SecurAccess with Chrome Enterprise provides access to internal web applications under the organization's existing security policies. In other words, UniconOS does not actually fix the damaged Windows installation. It gives employees another route to their applications while the IT department investigates or repairs Windows... [T]he installer shrinks the Windows volume, creates a new partition, copies UniconOS onto it, and registers the necessary boot entries. Windows remains the default operating system during ordinary use, although administrators can configure boot delays and fallback behavior...

The approach could prove useful following a widespread ransomware attack or another defective Windows update resembling the CrowdStrike incident of 2024. Recovery would happen independently on every compatible endpoint instead of requiring IT employees to physically handle thousands of computers.
Thanks to Slashdot reader BrianFagioli for sharing the news.