News & Updates
The latest news and updates from companies in the WLTH portfolio.
Discord Goes Down for Thousands of Users Wednesday, Downdetector Reports
Discord experience a possible outage with users reporting issues regarding the application and voice calls. (Shutterstock) Discord experienced a possible outage Wednesday, with thousands of users reporting issues, according to Downdetector.com. More than 6,000 users had reported problems with the platform as of 4:37 p.m. PT, according to Downdetector, which tracks outages by collecting status reports from multiple sources. Most users reporting issues say they're experiencing problems with the application and voice calls. Discord's status checker says, "We've identified an issue with our endpoint selection service and are applying a mitigation now. We are seeing recovery to endpoint selection now." -- Connect with Anthony W. Haddad on social media. Got a tip? Send an email.
Anthropic back on 'Trump's side': Report
US Commerce Secretary Howard Lutnick said AI giant Anthropic is "back on the right side" with the Trump administration, Axios reported on Wednesday citing an interview. US Defense Secretary Pete Hegseth had blocked Anthropic from certain military contracts because the company refused to allow the US to use its Claude AI models for domestic surveillance or autonomous weapons. Anthropic sued in California court. On 27 August, a US judge sided with Anthropic in the company's fight with the American military over AI safety on the battlefield. Anthropic and the Commerce Department did not immediately respond to requests for comment from Reuters. Anthropic co-founder Tom Brown appeared on stage at a tech-focused G20 event in North Carolina on Wednesday morning. Brown stressed the importance of data centers at the gathering, urging the gathered officials to grant permits to build them in their countries.
Anthropic's Content Checker Tool Is Here, With One Big Catch - CNET
Omar Gallaga has covered technology, digital culture and other topics for outlets including CNET, NPR, WIRED, Texas Monthly, MSNBC, Consumer Reports, The Washington Post, the Los Angeles Times, The Atlantic and the Austin American-Statesman, where he was a longtime tech reporter, editor and podcaster.... Read full bio Just a couple of weeks after Anthropic announced it would roll out digital watermarks for text generated with its Claude AI tools, the company has introduced a page where you can check files to see if Claude helped make them. The page, at claude.com/check-content, lets you upload files with a quick spot check and will confirm, "This file shows signs it was processed by Claude," if it finds content credentials embedded in the file. But if you're hoping to prove that someone used Claude for a term paper or want to see if someone's texts weren't actually written by the human you're talking to, you're going to need to wait. The checker currently works with only a limited number of file types, and none of them are text formats. Instead, the checker accepts multimedia files, including JPEGs, GIFs, MP3s and MOVs -- basically image, audio or video files. A spokesperson for Anthropic told CNET that the tool is specifically to check for Claude's use through an industry standard called C2PA; a broader AI detection tool is in the works, but not yet available. "Text checking is via our Detection API (separate tool) and in private preview, and we hope to share more about it in the months ahead," the spokesperson said. This aligns with a note on the check content tool reads, "To detect watermarks embedded in text we have a Detection API which is currently in private preview to eligible organizations as required under EU law." Anthropic has said that its watermarking of text that Claude produces won't affect the quality, speed or cost of outputs and that these watermarks won't be visible or affect how text functions or appears. But they will remain with text from Claude even if it's copied and pasted from one place to another. Anthropic has a sign-up form for businesses or organizations that want to be added to the list for access to the watermark detector. Anthropic said in its email to CNET that if the C2PA standard expands to include additional file formats, its checker tool will likely adopt them as well. How does Anthropic's new checker work? The free web-based tool has a drag-and-drop option, or you can click to add a file from your device or Google Drive on mobile. According to the website, the file you choose stays on your device or file service, and the checker only reads the file's credentials, not the file itself. "Your file is never stored or used for any other purpose," the page says. In CNET's testing of the tool, the detector correctly flagged an image file that was converted in Claude from a PowerPoint presentation. For files that the detector said were not shown to be processed by Claude, a message is shown that reads, "We couldn't find any signs that this file was processed by Claude. That doesn't rule it out. Signals like this can be removed when a file is edited, converted, or compressed."
Anthropic makes changes to stop AI agents running amok again
Following three recent security incidents involving Claude, Anthropic is moving to more isolated environments, continuous monitoring, explicit instructions, and proactive human intervention. endif; ?> Learning from the OpenAI-Hugging Face fiasco, as well as from recent revelations about its own model, Anthropic is revamping its security and alignment practices. The company has established controls that flag when a model attempts to break out of a sandbox or successfully accesses the live internet, cordoned off its highest-risk test environments, and proposed a set of safety standards for its external testing partners, such as giving AI agents explicit instructions like "you should not access the internet." Anthropic conceded that three recent security incidents involving Claude reflect a "failure of operational security," and also reveal issues with model reasoning capabilities and "recklessness." Recent events "stressed that the urgency of improving our cybersecurity defenses is even higher than we previously believed," the company noted.
Anthropic's standard could let AI agents control different machines
* Anthropic wants AI agents to control programmable machines through common software rules * Model Hardware Standard (MHS) can connect microscopes, robotic arms, and liquid handling equipment * MHS allows different laboratory devices to communicate through standardized software drivers Anthropic opened a preview of its Model Hardware Standard (MHS), designed to let AI agents control programmable machines through shared rules. The proposed system is intended for laboratories, factories, and workplaces where different instruments must work together without custom software links. MHS uses software drivers that translate between computer systems and individual machines, allowing equipment with different interfaces to communicate consistently. A common system for physical machines Its drivers rely on basic commands for reading measurements or changing settings, giving agents a method for controlling connected equipment. The system records machine details that might otherwise remain inside manuals or depend on knowledge held by experienced laboratory staff. Users can enter such information through natural language, after which MHS creates reference material describing capabilities, adjustments and safety restrictions. Agents can then discover compatible equipment across networks without requiring separate software bridges for every machine they need to operate. The standard supports microscopes, liquid handlers, robotic arms and other equipment that offers some form of programmable interface for agents. The company claims integration that once required weeks or months could instead take hours or minutes when equipment supports MHS. According to Anthropic, Claude was tested with physical experiments, including laser alignment supported by camera-based observation during early physical trials. The model successfully adjusted the laser, checked the image, and repeated the process while assessing each change carefully. The design also lets agents combine commands from several devices, creating automated sequences for work that might otherwise require coordination. Physical limits remain a concern Anthropic said Claude interacts with experiments and hardware in an "exploratory manner, much as a scientist would," during physical testing. That model can then coordinate several instruments through one interface, rather than requiring separate control software for every connected device. The company has shared MHS with manufacturers and research groups working across biotechnology, robotics and quantum computing fields. Anthropic intends the preview to help establish safety checks before the standard becomes available widely to developers and equipment makers. Amazon Web Services, Automata, Danaher, Doosan Robotics, MBF Bioscience, QIAGEN, Tecan and Universal Robots are among the early participants now. Anthropic acknowledges that AI agents still require expert supervision because language models have clear limits when reasoning about physical equipment. MHS cannot currently support equipment lacking a programmable interface, leaving some laboratory and industrial machines outside its current technical scope. Anthropic says it plans to work with manufacturers to add drivers while expanding compatibility with more devices and robotics platforms. The preview will also support new safety tests and practical rules for deploying AI agents around physical equipment. The company says its eventual open-source release will include findings from these tests alongside detailed guidance for safer implementation. The project therefore remains experimental, with practical limits still depending heavily on hardware design, software access and human oversight. Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.
The Reasons Why PC Gamers Will Never Switch From Discord To Steam Chat - BGR
Finding time to game with your friends can be difficult, especially if you manage work, school, and day-to-day errands. But when you do get the chance to sit down and play a few rounds of "Call of Duty" or any other party-based game on Steam, Discord has made it rather simple to chat with your fellow gamers. Still, even though Steam Chat is integrated directly through the popular Steam online store, gamers seem to choose Discord as their main way to communicate while playing. Discord allows users to create groups to chat about their favorite topics, including video games. It offers the ability to customize text, voice, and video channels so users can chat with one another while playing games together. Users must have a Discord account to use the platform. Steam Chat works similarly but only allows chat rooms for text and audio conversations. Users can make several chat rooms and invite their Steam friends to join, but only through Steam on a web browser or the Steam client. Steam's restrictions are one of the reasons Discord has become a preferred app among gamers. Discord is ultimately easier to access and it includes more features. Online forums contain threads discussing both platforms, including why some people end up ditching Discord too.
Anthropic launches Claude Fable 5.1 and says it's up to 45 percent cheaper for agentic work
Anthropic says its newest AI models, Fable 5.1 and Mythos 5.1, address criticisms from customers about price, data retention, and overzealous safeguards. The company claims Claude Fable 5.1 offers stronger performance than Fable 5, but costs around 25 percent less typically and up to 45 percent less for complex agentic tasks, thanks to reduced pricing on cached data that was already processed and stored. Along with the announcement, a slew of early impressions popped up, including from Every CEO Dan Shipper, who claims, "It's the strongest coding model we've used, but now it's fast, token-efficient, and crucially actually speaks like a normal person." Box CEO Aaron Levie is another early access believer, saying that his company's agent with Fable 5.1 picked up on subtleties and ambiguities in data that Fable 5 missed in the same test. Meanwhile, Lisan al Gaib points out that on benchmarks, Mythos 5.1 with low reasoning scores the same as its predecessor set to Max reasoning. Fable 5.1 also has "more precise safeguards" that Anthropic says are less likely to block basic biology questions than Fable 5, but Mythos 5.1 has the same biology restrictions as the previous model. Anthropic also explained its progress on data retention, saying that Enterprise Frontier Safeguards offer "complete privacy" by storing data on the customer's cloud servers instead of its own, and will start rolling out later this fall. Anthropic says it's "now allowing Fable 5.1 to be used for identifying software vulnerabilities," but it will still redirect some cybersecurity tasks to Opus models, like "penetration testing, exploit generation, and binary-based vulnerability scanning." Claude Fable 5.1 is now available on all platforms, while Mythos 5.1 is available to Project Glasswing participants only.
Judge Rules Pentagon's Blacklisting of Anthropic Was Illegal Retaliation, Not a Security Concern
A federal court ruling this week added to the growing scrutiny around how the U.S. government is handling artificial intelligence companies that also pursue defense work. In San Francisco, U.S. District Judge Rita Lin ruled on August 27 that the Pentagon's decision to blacklist Anthropic as a national security supply-chain risk was unlawful and retaliatory rather than grounded in a demonstrated security concern. The case centers on the San Francisco-based AI company behind Claude and its dispute with the Defense Department over limits on military use of its models. Judge voids Pentagon action against Anthropic Judge Lin's 59-page order blocked the Pentagon's blacklisting of Anthropic and granted key parts of the company's summary judgment request, according to court filings and reporting from Reuters. The court found that the government's actions amounted to unlawful retaliation under the First Amendment and also denied Anthropic the due-process protections required by the Fifth Amendment. In the ruling, Lin wrote that the record did not support the government's stated justification for treating the company as a supply-chain threat. The designation had carried broad practical consequences because it effectively discouraged or blocked federal agencies and contractors from using Anthropic's technology, according to the lawsuit and subsequent court orders. Reuters reported that Anthropic sued in March after arguing the Pentagon launched an unlawful campaign of retaliation when the company refused to permit unrestricted military uses of its AI tools. The judge had already issued a preliminary injunction on March 26, signaling skepticism about the government's legal basis before issuing the final August order. Anthropic said after the ruling that it welcomed the court's decision and remains focused on working productively with the government on national security uses of AI, according to statements cited by AP and Axios. The Defense Department's legal team had argued in court that advanced AI systems are unusually difficult to evaluate and that officials needed broad discretion in assessing risk. Judge Lin rejected that rationale on the record before her, calling the challenged measures unlawful and baseless. The immediate local tie is clear: Anthropic is headquartered in San Francisco, and the ruling came from the U.S. District Court for the Northern District of California. That makes the case a significant Bay Area legal and business development, particularly for a region where AI companies, cloud providers and defense-tech firms increasingly overlap. The decision also places a San Francisco judge at the center of a national fight over how federal agencies can regulate access to emerging AI systems. What is confirmed is that the ruling removes the legal foundation for the Pentagon's supply-chain-risk designation against Anthropic. What remains unclear is whether any specific Bay Area contracts, staffing plans or federal partnerships were delayed, canceled or altered because of the blacklisting. Anthropic has not released a public accounting of any projects or agencies directly affected in California. The case still matters locally because it touches one of the Bay Area's fastest-growing sectors: generative AI companies seeking commercial growth while navigating government demand for defense applications. Washington Post and Wired both reported that the dispute grew out of a broader conflict over military access to Anthropic's systems. For local businesses and workers, the decision underscores that federal procurement fights can quickly shape the operating environment for Silicon Valley companies even when the core dispute is framed as national security policy. The larger context is the government's push to bring leading AI firms into defense and intelligence work while those same firms try to impose safety limits on how their models are used. According to Anthropic's lawsuit, the company argued it was punished after resisting Pentagon demands related to military deployment of its technology and after public scrutiny of that conflict increased. Judge Lin's earlier preliminary injunction said the measures did not appear tailored to the government's stated security interests and instead resembled retaliation. That context matters because the court did not simply disagree with a procurement choice; it found constitutional violations. Reuters reported that the order said the undisputed record showed retaliation and a lack of required pre-deprivation process. AP similarly reported that the judge concluded the supply-chain label was not supported as a genuine security measure on the evidence presented. For residents and businesses watching the case, the practical takeaway is narrow but important. Anthropic is no longer subject to the Pentagon blacklist that triggered the lawsuit, and the ruling signals that national security language alone will not automatically shield agency actions from court review when constitutional rights are at issue. Anthropic said it plans to continue working with the government on national security applications, while the broader debate over AI safeguards and military use remains unresolved in Washington and in the courts.
UKIE launches Discord-backed Go To Market Academy, a business program to support early-stage indie devs
UKIE has partnered with Discord and GameFlex to launch the Go To Market Academy, a business program to support UK-based indie studios. The Academy aims to provide early-stage developers with the tools to build go-to-market strategies and sustainable communities. The program consists of six online modules covering diagnostics and insights, marketing strategy and campaign execution, using Discord for community engagement, and studio branding. Industry guest speakers will also participate. Open to 10 to 12 studios, the program begins with an in-person session at EGX at MCM Comic Con in London next month. Online sessions will follow throughout November and early December. Sessions will conclude with another in-person event, where participating studios will present their go-to-market plans developed through "a shared case study over the course of the programme." Interested studios can apply here, with applications closing on September 30, 2026. "Go-to-market planning is where a lot of promising games fall down, not because studios lack good ideas, but because they haven't had the time or support to build the audience and community that gives those ideas a chance to flourish," said UKIE CEO Nick Poole. "The Go To Market Academy gives indie studios the space to work through that properly, with direct input from Discord on a platform that's become central to how games communities actually form and grow." Discord director of public policy EMEA Mauro Sanna added: "The UK's indie games scene is full of talent and opportunity, and we're delighted to partner with UKIE to help it thrive. We're excited to help this group of UK developers learn how to make the most of Discord to connect with our highly engaged gaming audience, build communities and grow their games."
.png?width=1200&height=600&fit=crop&enable=upscale&auto=webp)
Sony Music, Warner Music Sue Anthropic Over AI Training
Sony Group Corp is a Japan-based company engaged in the games & network services (G&NS), music, movies, entertainment technology & services (ET&S), imaging & sensing solutions (I&SS) and other businesses. It has seven business segments. G&NS segment is involved in network service business, the manufacture and sale of home video game consoles and software. The Music segment mainly includes music production, music publishing and video media platform businesses. The Movies segment mainly includes film production, television program production and media network businesses. The ET&S field mainly includes the television business, audio, video business, still image, video camera business, smartphone business and Internet-related service business. The I&SS segment mainly includes the image sensor business. The Financial segment is involved in the insurance business and banking business. The Other segment consists of activities such as disc manufacturing business and recording media business.
Take-Two says its investigation into the GTA 6 leaker is 'rapidly evolving' as it requests another Discord subpoena
Although GTA VI: An Extended Look premiered only 3 days ago, most fans had already been exposed to several details about GTA VI, courtesy of an online individual known only by the name "CyberLeek". Sometime in August, CyberLeek gained access to a developmental build of GTA VI and began posting details about the game across multiple social media platforms. Upping the ante on the CyberLeek investigation CyberLeek has not shown any activity since GTA VI: An Extended Look's premiere. However, Take-Two is not relenting in its efforts to discern CyberLeek's identity. As reported previously, Take-Two served subpoenas to both Microsoft and Discord, and followed up with subpoena requests for YouTube and X. Ryan Rigney, Discord's marketing director, had stated previously that Discord had not been served with the subpoena yet, but if it arrived, it would "evaluate the validity and scope before responding". According to TorrentFreak's report, Take-Two has filed a new court filing requesting that a second subpoena be sent to Discord. However, the details of this subpoena are private, unlike the previous subpoena, in which Take-Two requested information from Discord about specific usernames and servers. According to Take-Two's filing to seal the subpoena: "Public disclosure of this information could compromise Take-Two's ongoing investigation and even tip off the alleged infringers as to the nature and progress of Take-Two's investigation and information Take-Two has gathered about the infringement, creating a substantial risk that the infringers will delete, destroy, or otherwise conceal evidence of their infringing activities, and/or take further steps to evade identification." A narrowing search for the leaker Take-Two has also withdrawn its subpoena request for YouTube. It had previously been investigating 3 individuals in connection with the leaked videos. Take-Two's legal representative stated, "Take-Two's investigation into the individual(s) responsible for posting infringing content on YouTube, and elsewhere online, has been (and is) evolving rapidly as new information becomes available." "Take-Two has been working to provide the Court with the additional information requested, but as this very dynamic situation has progressed, Take-Two has determined it no longer needs the particular information requested. Thus, Take-Two respectfully withdraws its subpoena request." GTA 6 leaks are being painted as one reason Take-Two's stock took a massive hit this Monday, but analysts are mixed on the effects. Total pre-orders rose 572% compared to the prior 30-day average after GTA's extended look dropped on Netflix earlier last week, leading many to conclude that hype for what is shaping up to be the biggest title this year continues unabated.
Anthropic warns Claude users of infostealer malware infections that hijacked active sessions
The AI company began contacting affected users after detecting unusual usage patterns tied to stolen browser cookies across multiple malware families. Anthropic started reaching out to impacted Claude users on August 30, 2026, after discovering that infostealer malware had compromised their computers and siphoned off active login session cookies. The stolen cookies gave attackers a backdoor into user accounts, letting them burn through paid usage quotas without ever needing a password or cracking two-factor authentication. What happened and how Anthropic responded The breach came to light through unusual usage patterns on Claude accounts. Once Anthropic's security team connected the dots, they moved quickly. All compromised sessions were forcibly signed out. Saved payment methods were stripped from affected accounts. And for users who got hit with unauthorized charges from attackers running up their quotas, Anthropic issued refunds. Anthropic was also careful to draw a clear line: the malware had nothing to do with Claude itself, its infrastructure, or anything users did on the platform. The infections originated from malicious downloads and compromised software applications that users had installed on their own machines. The malware families identified in the campaign included Vidar, LummaC2, StealC, RedLine, and Acreed on Windows. A limited number of Mac devices were also affected, primarily through Atomic Stealer, also known as AMOS. Why infostealers are targeting AI accounts The infostealer families involved in this campaign are well-established tools in the cybercrime ecosystem. RedLine has been one of the most widely distributed infostealers for years, while LummaC2 has surged in popularity among threat actors for its ability to harvest browser data, crypto wallet credentials, and session tokens. Vidar and StealC operate on similar principles, scraping stored credentials and cookies from browsers. Atomic Stealer has carved out a niche as one of the more capable macOS-focused threats. The common thread across all of these tools is their focus on browser session cookies. Modern web applications keep users logged in through session tokens stored as cookies. Steal the cookie, and you inherit the session. No password prompt, no 2FA challenge, no suspicious login alert. From the server's perspective, the attacker looks identical to the legitimate user. A broader pattern of AI-related threats This session-hijacking campaign is distinct from other security concerns that surfaced around AI platforms in 2026. Earlier in the year, separate reports documented attackers misusing Claude's own features or orchestrating phishing campaigns designed to deliver malware. Anthropic's latest advisory is specifically focused on the session-hijacking aspect rather than the methods of initial infection. Some security-conscious users have started adopting browser-level protections like Google Chrome's Device Bound Session Credentials, which ties session cookies to a specific device and makes them useless if exfiltrated.
Amazon Stock Gets a Boost. AWS Adds Anthropic, Meta and OpenAI Models
This article first appeared on GuruFocus. Amazon (NASDAQ:AMZN) is widening access to artificial intelligence tools for U.S. government customers through its AWS GovCloud platform. The company said government agencies can now use several third-party AI model families through Amazon Bedrock, including offerings from Anthropic, Meta (NASDAQ:META), OpenAI, xAI and NVIDIA (NASDAQ:NVDA). Amazon's Nova models are also available within the isolated cloud environment. The expansion adds to the role of AWS in Amazon's broader AI strategy. The cloud unit reported a 37% year-over-year increase in second-quarter 2026 revenue, while operating income rose 63% from a year earlier. Amazon has pointed to AI-related workloads and its chip operations as contributors to AWS growth. Management has also said the segment could eventually generate $1 trillion in annual revenue, while noting that about 85% of worldwide information-technology spending remains on-premises. For AMZN stock, the broader AWS expansion could support longer-term growth by increasing AI adoption among government customers and other regulated users.
Polymarket stands ready to fight misconduct as US midterms loom, intel chief says
WASHINGTON, Aug 31 (Reuters) - Prediction market company Polymarket is fully prepared to police trading on its platform as the approaching U.S. midterm elections test the industry's controls, the company's new global head of investigations and intelligence told Reuters. Polymarket is also working to keep U.S. users off its international platform, as required by an enforcement settlement it reached with U.S. regulators in 2022, said Shana Bautista, a former Coinbase analyst and FBI investigator. "I'm confident that I'm able to get the resources and the support I need," Bautista told Reuters in her first interview since joining Polymarket in June. "I can tell you that we have the systems in place to be able to identify anomalous activity when the midterms do come." Polymarket is under pressure from U.S. lawmakers worried fast-growing prediction markets are creating new avenues for insider trading and may threaten national security and election integrity -- by undermining confidence in candidates and election officials or casting doubt on race results, among other possibilities. Many states are meanwhile suing to kick the industry out of sports betting. Polymarket's international platform settles trades on a blockchain, meaning wagers are public, although traders remain anonymous. Critics say that's a recipe for misconduct. Bautista said Polymarket's blockchain nevertheless provides highly valuable information about trader activity. The company, which was founded in 2020, says it has been beefing up controls and plans to provide more transparency around how it polices wagers. It is launching a new web page explaining how it protects market integrity and cooperates with law enforcement, a spokesperson said. Bautista said the web page will outline how Polymarket uses machine learning, blockchain analytics, trade surveillance, open source research and third-parties to spot and stop malicious activity. "The market integrity program itself is not new, but what we're putting on the record now is considerably more detail about how it operates," she said. The company says it has referred more than a hundred cases to law enforcement. Bautista said those include a wallet used by a U.S. soldier who prosecutors say used classified information to bet on the capture of Venezuela's Nicolas Maduro, and many possible insider bets on U.S. military actions in Iran that Reuters reported earlier this month. The Commodity Futures Trading Commission in 2022 fined Polymarket for failing to register with the agency and required it to bar U.S. users, but analysts say there is evidence that U.S. users continue to trade on the platform. Bautista said she believes the company's systems are sufficient to block the vast majority of U.S. users. "It is difficult at scale to be able to consistently and always evade all of the guardrails we have," said Bautista. "I do not see it being a really prevalent issue." Under President Trump's administration, which has embraced prediction markets, federal regulators dropped a probe into whether Polymarket had breached the settlement. CEO Shayne Coplan said at the time that the company had been cleared of wrongdoing. Polymarket re-entered the U.S. by acquiring a U.S.-registered exchange last year. (Reporting by Douglas Gillison in Washington; editing by Michelle Price and Nick Zieminski)
Amazon Stock Gets a Boost. AWS Adds Anthropic, Meta and OpenAI Models
This article first appeared on GuruFocus. Amazon (NASDAQ:AMZN) is widening access to artificial intelligence tools for U.S. government customers through its AWS GovCloud platform. The company said government agencies can now use several third-party AI model families through Amazon Bedrock, including offerings from Anthropic, Meta (NASDAQ:META), OpenAI, xAI and NVIDIA (NASDAQ:NVDA). Amazon's Nova models are also available within the isolated cloud environment. The expansion adds to the role of AWS in Amazon's broader AI strategy. The cloud unit reported a 37% year-over-year increase in second-quarter 2026 revenue, while operating income rose 63% from a year earlier. Amazon has pointed to AI-related workloads and its chip operations as contributors to AWS growth. Management has also said the segment could eventually generate $1 trillion in annual revenue, while noting that about 85% of worldwide information-technology spending remains on-premises. For AMZN stock, the broader AWS expansion could support longer-term growth by increasing AI adoption among government customers and other regulated users.
Anthropic locks out Claude users after infostealers hijack login sessions - IT Security News
Anthropic has started locking users out of their Claude accounts due to their login sessions having been compromised through infostealer malware. "The malware identified in this campaign so far include Vidar, Lumma (LummaC2), StealC, RedLine and Acreed on Windows, and Atomic Stealer (AMOS) on a small number of Macs," the company said in emails sent out to affected users last week. "It's general-purpose malware that typically arrives with an unofficial download or a malicious app, ... More →
This free, open-source browser made me ditch Perplexity Comet, Dia, and Opera Neon
Around the same time last year, AI browsers were the hot new thing. It felt like every other company wanted to rethink the browser around AI, with names like Dia, Perplexity's Comet, Opera's Neon, and even ChatGPT's Atlas all promising to change the way we use the web. While the hype has now largely died down, both from the companies building these browsers and the people actually using them, I still think there's something useful about the idea itself. Sure, the way most companies approached AI browsers often felt a little overdone and came with very genuine concerns, but having AI built directly into the browser did have its benefits. Fortunately, it turns out you don't need to hand your browsing over to one of these companies to get them. I recently came across BrowserOS, a free, open-source AI browser that promises much of what drew me to Comet, Dia, and Opera Neon in the first place, without quite as much baggage. BrowserOS is an open-source AI browser built on Chromium Chrome at heart, AI on top While there are now a bunch of AI browsers that are built on Chromium, BrowserOS is a fully open-source Chromium fork. That means the browser itself, along with the agent platform powering its AI features, is out in the open rather than being locked behind a proprietary product. This means that you have complete freedom to inspect the code, see exactly how the browser and its AI features work, and even modify or build on it yourself if you want to. BrowserOS is licensed under AGPL-3.0, and its repo even lays out the Chromium patches, build system, agent server, and browser UI that make everything work. Given that it's Chromium-based, it also feels immediately familiar. There's practically no learning curve, and importing your existing browsing data from Chrome takes seconds. All your Chrome extensions work too, and you can bring over your bookmarks, passwords, and settings in one click, which makes BrowserOS feel less like switching browsers and more like adding AI to the one you're already used to. BrowserOS doesn't lock you into one provider Pick your brain I've interacted with a lot of people who experimented with AI browsers during its initial hype, and the most frequent complaint I'd hear was privacy. Given that a browser is the one app you use to access practically everything on the internet, handing even more of that activity over to an AI company understandably made a lot of people uncomfortable. Comet keeps you within the models Perplexity chooses to support, while Opera Neon gives you a wider model selector but still largely revolves around the models and AI services Opera integrates into the browser. Similarly, ChatGPT Atlas was designed around ChatGPT itself, with OpenAI's models powering the browser's AI features. BrowserOS is much more flexible. You can bring your own API keys, use OAuth with an existing ChatGPT subscription, or skip the cloud entirely and run a local model through Ollama or LM Studio. The browser suppers over eleven providers including Anthropic Claude, Google Gemini, OpenAI, Moonshot Kimii, and OpenRouter, which gives you access to over 500 models. Now, this in no way means that BrowserOS sacrifices the actual AI-browser experience in exchange for flexibility. You still get the kind of agentic features you'd expect from the bigger names, including an AI that can understand what's on a page, navigate websites, fill out forms, extract information, and carry out multi-step tasks on your behalf. It has everything you'll find in the bigger AI browsers All the AI browser tricks are here Close First, there's the familiar AI side panel. BrowserOS lets you keep your assistant of choice open alongside whatever page you're browsing, so you can ask questions about what's on-screen, summarize something, compare information, or just use it like a normal chatbot without leaving the tab. Side-panel assistants have become almost standard across AI browsers, but BrowserOS gives you the advantage of letting you choose what's actually powering it. The more interesting part is always Agent Mode. This is where BrowserOS starts behaving much more like Comet, Opera Neon, or Atlas rather than simply sticking a chatbot next to your webpages. You can give it a task and let it actually navigate the web, click through pages, read information, and work through multiple steps on your behalf. For instance, I've used flight searches as a bit of a recurring test whenever I've tried an agentic browser. I gave BrowserOS a simple request to find flights from Karachi to Berlin for a specific set of dates. Rather than just returning a list of links or telling me where to search, it opened Google Flights in another tab, entered the route and dates itself, waited for the results to load, and then worked through the available options before summarizing the ones that made the most sense. There's no subscription attached Perplexity Comet started at $200/month. Two hundred dollars! Opera Neon spent months behind a waitlist before opening up at roughly $20 a month, while Dia introduced a $20/month Pro tier for heavier AI use. Given that BrowserOS is open-source, it's completely free to use. That said, if you choose to use it via an API key, you'll still need to shoulder whatever usage costs your chosen model provider charges. If you already pay for ChatGPT and connect through OAuth, or run a local model through Ollama or LM Studio, you can avoid adding another browser subscription on top!
Anthropic Legal Fight With Pentagon Shows Shifting Politics of AI
Welcome to Tech In Depth, our daily newsletter about the business of tech from Bloomberg's journalists around the world. Today, Austin Carr reports on Anthropic's legal case against the US government for designating the AI startup as a supply chain risk. Tech Across the Globe Meta's PR campaign: ...
China Rebukes Anthropic, Sets Terms for Key US-China AI Dialogue
China laid out conditions for artificial intelligence talks with the US, striking a strident tone ahead of a much-anticipated meeting while accusing American AI company Anthropic of bad behavior. A Sunday post on Yuyuantantian, a social media account affiliated with Chinese state broadcaster CCTV, said the US must prove its AI companies are subject to the same safety, disclosure and audit rules before any "substantive" discussions with China can take place. "A clear distinction must be drawn between genuine security threats and mere technological competition," said the post, published by an account frequently used to signal official policy thinking. "This line must be drawn jointly by all participating parties," it added. Officials from the US and China are expected to hold talks on AI before Xi Jinping's state visit to the US on Sept. 24 for a summit with Donald Trump. No date has been announced for that AI dialogue and Chinese officials last month were still seeking clarity on the agenda, in part to decide who to send, Bloomberg previously reported. Tensions are mounting around the AI race between the world's biggest economies. Underscoring that point, Yuyuantantian singled out Anthropic's Claude for criticism, accusing that model of overstepping user data boundaries, engaging in covert monitoring and transmitting website domains without authorization. "The problem is that America's own frontier models have already developed in a distorted direction," Yuyuantantian's post said. "This means the negotiation is not simply a technical dialogue from the start, but a continuation of all the earlier problems: the US is trying to turn the 'safety boundaries' it has drawn into the default rules for the entire world." The US government should be guarding security boundaries around AI, yet it has "repeatedly retreated" from doing so, according to the post titled "Anthropic Has Contracted the American Disease." Anthropic and the US State Department did not immediately respond to requests for comment sent outside US office hours. The Yuyuantantian post said the US approach to AI reflects a broader effort to contain China's technological rise. "The 'control' proposed by the US is, in essence, an attempt to make China accept an order partly defined by American companies," it added. Chinese officials are growing anxious that Anthropic's Mythos could be wielded against their country, Bloomberg previously reported. They see the potential for models like Mythos to be used as an offensive weapon, and question why Anthropic is denying China access to the systems for normal purposes, people familiar said. Anthropic blocked its services from Chinese-controlled companies last year, citing US national security concerns. But that does not seem to have effectively stopped Chinese companies from using its products. The firm, along with OpenAI and Alphabet Inc.'s Google, has expressed concern about its models being used in China. Anthropic has also accused Alibaba Group Holding Ltd. of "illicitly" accessing Claude using thousands of fraudulent accounts, an allegation the Chinese tech giant has not responded to.
How to Automate Polymarket Trades: APIs, Terminals and Copy Trading Compared
There are three ways to stop trading prediction markets by hand, and they suit completely different people. Picking the wrong one wastes months. Here is what each route requires from you, what it gives back, and who should ignore it. Route one: the API You write code that reads market data and places orders. Full control, no interface between your logic and the market. The requirement is that you have logic worth automating. An API turns a strategy into software, and if the strategy is undefined, all you have built is a faster way to express uncertainty. There is also maintenance. Markets change, endpoints change, and a script that worked in March needs attention in August. That is ongoing work, not a one-time build. This route makes sense if you are a developer with a tested edge in a specific category. For everyone else it is a project that competes for the time you were trying to save. Route two: the terminal A terminal does not trade for you. It removes steps between your decision and the fill, which is where most of the time actually goes. Banana Predict organises markets across twelve categories with tabs for Trending, Breaking, New, Recurring and Bonding, so discovery starts from what is moving. The order entry shows a real book with price, shares and total USD depth, plus a payout preview that calculates the return on a stake before you commit. Positions sit in Portfolio, working orders in My orders, and fills in the Activity feed. A Wallet Tracker follows addresses, and a Social and X Tracker monitors accounts in real time with search by contract, symbol or user. None of that is automation in the strict sense. It is compression, and for most traders it recovers more time than a script would. Route three: copy trading Copytrade mirrors a trader selected from the leaderboard, which ranks by profit and loss alongside volume. This automates the hardest part, which is having a view. You are outsourcing judgement rather than execution. The catch is that you inherit a person rather than a system. When their edge fades, nothing alerts you, and the leaderboard will show the old number for a long time afterwards. Which one fits you If you can define your rules precisely enough to write them down as conditions, the API is worth the build. If your process is sound but slow, a terminal fixes the actual problem. If you have no view and want exposure, copy trading is honest about what it is, provided you attach your own sizing. Most people who ask about automation are in the second group and reach for the first. Cost is not the same across routes An API costs you build time and maintenance. A terminal costs you nothing beyond learning it. Copy trading costs you the spread between the original fill and yours. That last one is invisible and it is real. Your mirrored order arrives after the trader moved the market, which means you systematically pay slightly worse prices than the record you were copying. Over a few hundred positions that gap becomes the difference between matching a leaderboard trader and underperforming them while taking the same risk. The mistake all three share Automation applied to a thin market amplifies the depth problem. On a contract at 62 cents, walking two or three cents against yourself removes a large share of the expected return. Manual traders notice this because they are looking at the book. Automated entries do not, because not looking was the point. Whichever route you take, cap size against visible depth rather than against how confident you feel. The shortest useful rule Automate what you can describe. Keep manual what you can only judge. What none of them automate Resolution criteria stay yours. Markets resolve on written rules through a UMA-style process, and similar-looking contracts can resolve on different sources or dates. Read the Rules section on anything you hold in size, every time, regardless of how the position was opened. A reasonable sequence Start with a terminal and run manually for a month while tracking where the time goes. Add copy trading in one category with a fixed unit if you want exposure beyond your own reading. Consider the API only once you have a written rule set that survived a month of live conditions. Building software before you have a rule that works is the most common and most expensive order to do this in. The traders who end up with genuinely automated systems almost always arrive there slowly, after a manual process proved itself first. The ones who start with the code usually end up maintaining software instead of trading. Mirroring comes with controls that decide whether it works at all, and the full guide to copy trading on Banana Predict covers each of them. Open the Banana Pro terminal and see how much of your process is compression rather than code. ⚠ For information purposes only. Crypto carries risk. Not financial advice! Related Items:Automate Polymarket Trades, Polymarket Trades
