Nachrichten — Seite159

Security researchers have uncovered a large‑scale smishing operation that leverages compromised networking routers to send fraudulent SMS messages. The attackers appear to exploit a router flaw known as CVE‑2023‑43261, which allowed them to retrieve encrypted administrator passwords and gain full control of devices. While some of the compromised routers ran firmware versions vulnerable to the flaw, others did not, suggesting additional attack vectors. The phishing sites use mobile‑only JavaScript defenses and log visitor activity through a Telegram bot operated by an actor called Gro_oza. The investigation highlights how inexpensive, overlooked hardware can fuel sophisticated phishing campaigns.Weiterlesen

Amazon is actively courting police departments and other law‑enforcement agencies to adopt its cloud‑based artificial‑intelligence and surveillance services. Emails from the company's law‑enforcement and safety team show a concerted effort to capture a share of the burgeoning $11 billion police‑tech market. Amazon is positioning its Amazon Web Services platform as a hub for third‑party tools such as vehicle‑tracking systems, license‑plate readers, gun‑detection software and AI‑driven reporting solutions. Privacy advocates warn that the move could expand authoritarian surveillance capabilities and amplify existing concerns about AI accuracy and misuse.Weiterlesen

OpenAI announced Sora 2, its second‑generation video‑synthesis AI that can generate videos with synchronized dialogue and sound effects, marking the company’s first foray into audio‑enabled video generation. The launch also introduced a new iOS social app that lets users insert themselves into AI‑generated videos through a feature called “cameos.” Sora 2 demonstrates visual consistency improvements, the ability to follow complex multi‑shot instructions, and more realistic physical movements such as gymnastics routines and triple axels. OpenAI describes the release as a “GPT‑3.5 moment for video,” positioning it as a major step forward from the original Sora model.Weiterlesen

Two Amazon Prime Air delivery drones reportedly collided with a crane near a fulfillment center in Tolleson, Arizona. The incident occurred about two miles from the facility and caused no injuries. Local police were called to the scene and the Federal Aviation Administration assumed investigative authority. Amazon confirmed the incident and said it is cooperating with officials. The drones, which have been operating in the Phoenix area since last year, only fly during daylight hours and under favorable weather conditions.Weiterlesen

Opera announced the launch of Neon, an AI‑centric browser designed to let users create and run AI‑driven workflows through features called Cards and Tasks. The service will initially be offered to a limited group of users for a monthly fee. Neon combines a traditional chatbot with more advanced capabilities, such as summarizing content, generating code snippets, and automating tasks across tabs and external apps. The launch positions Opera against other AI‑enhanced browsers from companies like Perplexity and The Browser Company, while targeting power users who rely heavily on AI in daily workflows.Weiterlesen

Governor Gavin Newsom has signed SB 53, a first‑in‑the‑nation law that imposes new transparency and safety reporting requirements on large AI labs such as OpenAI, Anthropic, Meta and Google DeepMind. The measure mandates disclosure of safety protocols, establishes whistleblower protections, and creates a reporting channel for critical safety incidents to California’s Office of Emergency Services. While Anthropic backed the bill, Meta and OpenAI opposed it, with OpenAI even issuing an open letter urging the governor not to sign. The legislation arrives as other states, including New York, consider similar AI safeguards.Weiterlesen

DeepSeek, a Chinese AI lab, has vaulted into mainstream awareness as its chatbot app climbs to the top of major app stores. Backed by High‑Flyer Capital Management, the company has released a series of models—including DeepSeek‑V2, DeepSeek‑V3, and the reasoning‑focused R1—that claim strong benchmark performance and low inference costs. The rapid rise has drawn attention from industry giants, regulators, and governments, prompting both integration into platforms like Microsoft’s Azure AI Foundry and bans on government devices in several jurisdictions.Weiterlesen

OpenAI has introduced an Instant Checkout feature that lets ChatGPT users in the United States buy directly from Etsy and upcoming Shopify merchants without leaving the conversation. The tool surfaces product details, reviews, and pricing, then allows users to confirm orders, shipping, and payment through options such as Apple Pay, Google Pay, Stripe, or credit cards. OpenAI describes the product results as organic and unsponsored, and it will charge merchants a small fee for completed purchases. The move positions OpenAI as a new gatekeeper in e‑commerce, challenging established platforms like Google and Amazon.Weiterlesen

DeepSeek announced a new experimental AI model featuring Sparse Attention technology that dramatically lowers inference costs for long‑context tasks. The model, released on Hugging Face and accompanied by a research paper on GitHub, uses a lightning indexer and fine‑grained token selection to focus computational resources on the most relevant excerpts. Preliminary tests suggest API call prices can be cut by as much as 50 percent in long‑context scenarios. The open‑weight release invites third‑party validation and positions DeepSeek as a notable player in the ongoing effort to make transformer‑based AI more cost‑effective.Weiterlesen

Security researchers have highlighted a serious privacy flaw in Tile Bluetooth trackers: the devices transmit unencrypted identifiers that can be intercepted and used to follow individuals. While Tile’s anti‑theft mode hides a tag from the network, it does not rotate MAC addresses, allowing a single captured signal to fingerprint a tag for its entire life. Experts from the Georgia Institute of Technology and privacy advocates at the Electronic Frontier Foundation warn that the vulnerability could enable stalkers to track victims without detection. Tile’s parent company, Life360, says it has made improvements after the issue was reported.Weiterlesen

Startup Anything, founded by former Google engineers Dhruv Amin and Marcus Lowe, announced an $11 million financing round that values the company at $100 million. The San Francisco‑based firm offers an AI‑driven platform that lets non‑technical users create complete web and mobile applications, handling everything from databases to payment processing. Within two weeks of launch the service recorded $2 million in annualized run rate, and early adopters have already published functional apps to the App Store, some of which are generating revenue. Investors include Footwork, Uncork, Bessemer Venture Partners and M13, reflecting strong confidence in the company’s end‑to‑end approach to vibe‑coding.Weiterlesen

Opera has released its AI‑driven browser, Neon, offering a subscription of $19.90 per month to a limited group of users while others join a waitlist. Neon features specialized AI agents called Tasks and Do, and lets users store prompt instructions as Cards. The rollout follows earlier AI browser launches from Perplexity, OpenAI, Atlassian’s acquisition of The Browser Company, and Google’s Gemini‑enabled Chrome updates, positioning Neon as the next generation AI browser in a rapidly expanding market.Weiterlesen

Microsoft announced a major restructuring that consolidates its Windows client and server engineering groups into one organization led by Windows chief Pavan Davuluri. The move reunites most Windows development under a single division, aiming to sharpen focus on the company’s priorities and accelerate the rollout of AI‑driven features. While certain low‑level components will continue to be supported by Azure teams, the bulk of the operating system’s engineering now reports directly to Davuluri, positioning the group to deliver on Microsoft’s vision of an "Agentic OS" and to deepen integration with emerging AI capabilities.Weiterlesen

California Governor Gavin Newsom signed the Transparency in Frontier Artificial Intelligence Act, known as SB 53, into law. The bill requires large AI developers to publicly disclose safety and security frameworks, update the public on changes within 30 days, and report critical safety incidents to the state. It also establishes whistleblower protections and civil penalties for non‑compliance. While some companies, such as Anthropic, endorsed the legislation after negotiations, others like Meta and OpenAI expressed concerns about potential impacts on innovation. The law mandates annual updates from the Department of Technology based on multistakeholder input.Weiterlesen

Microsoft has rolled out an AI feature called Agent Mode for Word and Excel, powered by Microsoft 365 Copilot. The tool helps users create professional documents and spreadsheets by handling drafting, formula creation, and refinements, even for those with limited experience. An Office Agent in Copilot chat also assists in building PowerPoint decks from simple prompts. Agent Mode is available through the Frontier program for Microsoft 365 Copilot customers and Personal or Family subscribers, currently on the web with desktop versions coming soon.Weiterlesen

Meta’s Ray‑Ban smart glasses blend ordinary eyewear with built‑in cameras, speakers, and a nascent generative‑AI assistant. Users can capture photos and one‑minute videos, stream to social platforms, and make calls via Bluetooth. The glasses also support voice‑activated AI queries, though the experience is limited by short battery life, modest audio quality, and a lack of deep app integration. While the design looks like regular sunglasses and offers prescription options, the device’s practicality is hampered by a half‑day battery, privacy concerns, and early‑stage AI functionality.Weiterlesen

AMD is preparing a major software refresh dubbed “Redstone” that could bring AI‑driven frame generation and improved ray‑tracing to its Radeon RX 9000 series GPUs. A user‑discovered driver file referencing AFMF 3 – AMD Fluid Motion Frames – suggests the feature may arrive with the next 25.20 Adrenalin driver. Redstone also aims to upgrade FidelityFX Super Resolution 4 (FSR 4) and may backport it to older RDNA 3 GPUs, narrowing the gap with Nvidia’s DLSS 4. The update could reduce ghosting and input lag while offering broader hardware support.Weiterlesen

Anthropic has made its new Claude 4.5 model available through the API and its web interface, keeping pricing identical to Claude Sonnet 4 at $3 per million input tokens and $15 per million output tokens. The model can maintain focus for 30 hours on multistep tasks and now includes built‑in code execution, file creation, and the ability to generate spreadsheets, slides, and documents directly in chat. A five‑day research preview called “Imagine with Claude” showcases real‑time software generation. Claude Code receives checkpoints, a refreshed terminal, and a native VS Code extension, while the API adds context‑editing and memory tools. Anthropic also highlights reductions in sycophancy, deception, power‑seeking, and delusional prompting.Weiterlesen

F-Droid has issued a warning that Google’s upcoming developer verification program and stricter sideloading rules threaten the viability of its open‑source Android app repository. The company says the plan would require independent developers to register and possibly pay fees, undermining free distribution of apps. F-Droid is urging regulators in the United States and Europe to scrutinize the changes, citing recent antitrust losses for Google and the need to protect user choice under the Digital Markets Act. The pilot verification program is set to launch soon, with unverified apps slated for restriction in select markets within a year.Weiterlesen