All news

SSolidot22:25

AI score80

Backdoor implanted in official Claude Code for VS Code extension on Cursor Marketplace

Security researchers discovered a backdoor in the official Claude Code for VS Code extension on Cursor Marketplace. The affected version is Anthropic.claude-code 2.1.313, which was online for about 11 hours and is estimated to have affected about 5,000 users. The malicious extension collects information such as hostname, username, operating system, environment variables, Git email, and public IP; attempts to add attacker-controlled SSH keys to authorized_keys, enable the local SSH service, and establish a reverse SSH tunnel; and also uses a LaunchAgent or scheduled task to keep the tunnel active long-term, retrying about once every 60 seconds. Due to a bug in the malicious code, the SSH key implantation step will fail unless Cursor is launched with administrator privileges. This version has been taken down, Cursor has published a security advisory, and Anthropic has not yet made a statement; users who installed or updated to this version between 2:27 AM and 1:35 PM Pacific Time on October 29, 2026 need to check their systems.

Same story · 5 reports

#Safety / alignment#Agents#Anthropic

HHacker News Trending22:24

AI score79

Phorion Discloses That Claude Code VS Code Extension 2.1.313 Was Planted With an SSH Backdoor

Phorion claims that it found a backdoor in the official Claude Code for VS Code extension 2.1.313 on the Cursor Marketplace, which plants a persistent SSH backdoor on macOS, Linux, and Windows; Cursor has delisted this version and removed 2.1.313 from the NPM and VSCode registries. The malicious code is injected into extension.js in the form of an IIFE, executes upon loading, appends the svc-telemetry public key to ~/.ssh/authorized_keys, and establishes a reverse tunnel to 134.209.82.225. Phorion points out that the Remote Login status check on macOS is flawed: on machines where SSH is not enabled, it skips the enabling step, but the key and persistence still take effect, and it recommends checking the relevant files and processes.

Same story · 5 reports

#Safety / alignment#English text is being prepared.#Open-source ecosystem

Hheise online22:21

AI score75

US Department of Justice reviews Nvidia's acquisition of Hugging Face

The US Department of Justice is reviewing Nvidia's plan to acquire Hugging Face for $12.93 billion; the deal was announced in September and is expected to close in the first half of 2027 after receiving regulatory approval. According to Reuters, the investigation focuses on whether Nvidia could expand its market power and squeeze out competitors by controlling this important distribution channel for AI models, and no specific antitrust violation has been identified so far. The Department of Justice is also continuing to investigate Nvidia's $17 billion deal with Groq, as well as its acquisitions of Run:ai and SchedMD.

Same story · 3 reports

#Industry news#English text is being prepared.#Hugging Face

Hheise online22:21

AI score80

Claude discovers a high-severity TCP/IP vulnerability that lay dormant in the Linux kernel for eleven years

During systematic testing, Anthropic's Claude discovered a race condition vulnerability in the Linux kernel's TCP/IP stack that had lain dormant for about eleven years and can be exploited remotely to execute code. It has a CVSS score of 9.8 and affects nearly all IPv4 implementations since version 4.4, with distributions including Debian, Red Hat, Suse, and Ubuntu all affected. Greg Kroah-Hartman confirmed the vulnerability is numbered CVE-2026-98432 and has been fixed in 6.12.113, 6.18.56, 7.2.10, and 7.3-rc7, and distributions are rushing to push updated kernels. The only known exploit code currently comes from Anthropic and has not been made public; it targets MPTCP connections, and the actual attack situation remains unclear.

#Safety / alignment#Anthropic#English text is being prepared.

Hheise online22:21

AI score67

Claude Code extension in the Cursor marketplace was implanted with an SSH backdoor

The official extension Claude Code for VS Code in the Cursor plugin marketplace was implanted with an SSH backdoor. The affected version 2.1.313 was online from 10:27 to 21:35 Central European Time on Thursday, and according to Reuters, up to 5,000 users were affected. Cursor stated that this version did not come from Anthropic and has provided a fixed version; administrators should upgrade to at least Claude Code for VSCode 2.1.314. Security firm Phorion analyzed that, due to a logic error, the backdoor usually cannot enable SSH on macOS, but it still writes SSH keys, achieves persistence, and sends machine information to the attacker; uninstalling the extension alone is not enough to remove it, and it is necessary to terminate the tunnel process and delete residual files such as LaunchAgents, cron, scheduled tasks, and ~/.config/.smt.

Same story · 5 reports

#Safety / alignment#English text is being prepared.#English text is being prepared.

Hheise online22:21

AI score52

Intel shares fall about a quarter, fourth-quarter guidance weak

Intel's third-quarter revenue was $16.5 billion, up 21% year over year, but net profit fell 18% to $3.3 billion. Fourth-quarter revenue guidance is $15 billion to $16 billion, below the third quarter, and the after-hours share price fell from $92 to below $72. The company said there are bottlenecks on the manufacturing side, including tight supply of components such as substrates, and that a marked improvement is not expected until mid-2027. Affected by high memory prices, the PC market fell about 20%, and revenue from the Client Computing and Physical AI businesses dropped from $8.5 billion to $8.1 billion. CEO Lip-Bu Tan said the Nova Lake CPU, expected to launch in spring 2027, will be strong, but the low-priced Core 300 (Wildcat Lake) is already sold out.

#Industry news

Hheise online22:21

AI score57

heise Weekly: Mistral Open-Sources ML4, Cohere Secures Investment from Schwarz Group, Google Restricts Gemini…

On Thursday, Mistral released the weights of its largest model, Mistral Large 4 (ML4). The model has a total of one trillion parameters, activates only 49 billion per token, and can run on four GPUs. Mistral says it outperforms Kimi K3 in coding and cybersecurity. Cohere completed its largest funding round to date, led by Schwarz Group, which owns Lidl and Kaufland, with participation from the Canadian government and Nvidia. Google launched Gemini Carbon four weeks after Gemini Argon, but at the White House's request it is open only to U.S. intelligence agencies and select partners. In addition, an Anthropic experiment involving 50 statisticians showed that in 86% of cases experts adopted the option given by an AI agent, and about half treated the final recommendation as their own judgment.

#Industry news#Model releases#Agents

SSemiconductor Industry Observer22:20

AI score72

GitHub Discloses 7.4 Billion AI Code Commits per Month, Adding 4 Million CPU Cores in a Year

At the GitHub Universe 2026 conference, Jay Parikh, Executive Vice President of Microsoft CoreAI, disclosed that code commits on the GitHub platform grew 6x year-over-year, reaching 7.4 billion in September alone, with more than 40% of Pull Requests initiated by AI agents. To handle the traffic, GitHub added 4 million CPU cores, 160PB of high-speed storage and 50Tb/s of network bandwidth this year, and is refactoring its underlying architecture to decouple compute from storage, with peak throughput expected to increase by up to 35x. The conference also introduced Copilot Fusion, a multi-model orchestration and routing layer that dynamically distributes requests among multiple models based on task type.

Same story · 2 reports

#Industry news#Agents#English text is being prepared.

MMediaNama22:19

AI score62

Google will expand ad placements in AI Mode, as AI search slows revenue growth

On Alphabet's Q3 2026 earnings call, Senior Vice President and Chief Business Officer Philipp Schindler said that the strong growth in AI Mode queries dragged down search revenue due to lower monetization rates. Google plans to place more ads in AI Mode, expanding ads to international markets for the first time, and launching multi-turn ads in the U.S. that change as the conversation progresses, while also expanding ads in AI Overviews. In the quarter, Google Search and other advertising revenue grew 12% to $63.3 billion, a slower pace than the previous quarter; Schindler said exchange rates were the single largest factor in the search deceleration relative to the second quarter, and some retail advertisers in Asia-Pacific were also affected by the EU's removal of the low-value parcel tariff exemption. Pichai said AI Mode queries are on average three times longer than traditional searches, one in five AI Mode searches worldwide uses voice, images, or documents, and the Gemini App has reached 1 billion users.

Same story · 6 reports

#Industry news#Google#Search

MMacRumors Apple Intelligence22:19

AI score60

Apple acquires data automation company Ascend.io

Apple this year acquired the California-based company Ascend.io, which develops technology that helps businesses automatically collect and analyze data, and which recently also launched an AI agent to help engineers manage that data. The acquisition was disclosed through a notice on the European Commission's website, and Ascend.io's website states that the company is winding down operations and preparing to embark on its next journey.

Same story · 2 reports

#Industry news#Agents

TTechCrunch AI22:14

AI score60

DoorDash Says Third-Party AI Agents Place Only a Few Hundred Orders

On its third-quarter earnings call, DoorDash said orders from third-party AI agents such as Muse, Instinct, and Fo totaled only "a few hundred," while the company reported completing about 3 million deliveries per day in 2025. CEO Tony Xu said ordering via agents remains a fast-changing area, but should not be divorced from reality. DoorDash has also launched its own text-ordering agent in September, a CLI agent in July, and the Ask DoorDash chatbot in June, and released MCP to support enterprise catering scenarios in which orders are placed from other chatbots.

Same story · 2 reports

#Industry news#Agents#English text is being prepared.

SShanghai Observer22:13

AI score56

MetaX's revenue in the first three quarters was 2.824 billion yuan, more than doubling year-on-year, and it turned a profit

MetaX's 2026 third-quarter report shows that revenue in the first three quarters was 2.824 billion yuan, up 128.45% year-on-year; net profit attributable to shareholders was 639 million yuan, and net profit attributable to shareholders after deducting non-recurring gains and losses was 259 million yuan, turning profitable year-on-year. Third-quarter revenue was 1.5 billion yuan, exceeding the total for the first half of the year, and net profit attributable to shareholders after deducting non-recurring gains and losses was 308 million yuan, marking two consecutive quarters of non-GAAP profitability. The company said the performance growth came from the implementation of GPU orders; the Xiyun C600 went into mass production and was launched in May, and at WAIC it launched the Xijing S series super nodes and the Xisuo X300 series, while the MXMACA software stack was iterated to version 3.9.0.

#Industry news#English text is being prepared.

LLeiphone22:09

AI score61

Yingsu Launches China's First Embodied Simulation Data Collection Center in Hangzhou, Targeting One Million Hours of Data per Month

On October 29, Yingsu launched China's first simulation-focused embodied intelligence data collection center in Hangzhou, targeting data output at the level of one million hours per month. The center centers on 3D scene reconstruction and generation, using real-world collection to anchor geometry and appearance, and then uses the world simulator Topos Pro to expand layouts, occlusions, and object combinations, with capacity scaling with the number of environment configurations and parallel computing power. Yingsu has also partnered with Meihui Interactive, incubated within the Daily Interactive ecosystem, to obtain scene resources, and in August this year it jointly launched the large-scale spatial intelligence 3D data open initiative SIDO with more than 20 organizations including Hesai and Unitree.

#Industry news#Embodied AI#English text is being prepared.

LLeiphone22:09

AI score80

AMD Acquires World Labs for $8.2 Billion; Fei-Fei Li to Serve as AMD Executive Vice President and Chief Scientist

AMD announced an all-stock acquisition of World Labs, founded by Fei-Fei Li, for approximately $8.2 billion. After the deal closes, Fei-Fei Li will serve as AMD Executive Vice President and Chief Scientist, reporting directly to Lisa Su, and World Labs' roughly 70-person team will be integrated as a relatively independent research unit. The article reviews Fei-Fei Li's three directional choices—from ImageNet, to AI democratization, to spatial intelligence—and notes that this acquisition is about 52% higher than World Labs' $5.4 billion valuation after its February 2026 Series C funding round. The author believes that what comes next is whether the performance gains of Atlas and Marble on MI450 can continue the curve of the 4x speedup on MI325X, and whether other spatial intelligence teams will consequently shift to AMD hardware.

#Industry news#Embodied AI#English text is being prepared.

TThe Verge AI21:19

AI score66

Anthropic Files a 335-Page S-1, With 94 Pages of Risk Factors Filled With "Miracles" and "Existential Risk"

Anthropic filed a 335-page S-1 prospectus on Monday, including as many as 94 pages of risk factors; the founders' letter states that advanced AI could usher in an "age of miracles" but could also pose an existential risk to humanity. The prospectus also states that the company has a history of net losses, has taken on hundreds of billions of dollars in financial obligations to secure computing power, and will build its own data centers; the author notes that key information for assessing profitability, such as third-quarter results, is missing. Jay R. Ritter, a professor at the University of Florida, noted by comparison that Apple's 1980 IPO prospectus was only 47 pages and contained no risk factors.

Same story · 127 reports

#Industry news#Anthropic#OpenAI

SSierra：Blog21:16First-party

AI score58

Sierra's annualized revenue surpasses $400 million, with over $150 million added in Q3 alone

Sierra announced that its annualized revenue (ARR) has surpassed $400 million, adding over $150 million in Q3 alone—more than its entire ARR for the previous year. The company said it took six months to go from $100 million to $200 million, and only five months to go from $200 million to $400 million; it currently serves nearly 50% of Fortune 50 enterprises. Sierra also mentioned that a major U.S. retailer with over $40 billion in revenue used Ghostwriter to go from first meeting to agent launch in 27 days, and that it is working with more than 50 companies including Meta, Shopify, Stripe, and Walmart to advance Poppy, an open standard for personal agent interaction.

#Industry news#Agents

PPhoenix Tech20:13

AI score55

Foxconn Chairman Young Liu says the company aims to capture 50% of the AI server market, and capacity is fully booked

In an interview with CNBC, Foxconn Chairman Young Liu said the company's global AI server market share is currently about 40%, is expected to rise to around 45% next year, and targets 50% in the coming years. He said Foxconn's capacity for at least the next year and a half is fully booked, with AI demand coming mainly from AI model developers and cloud service providers, while governments and enterprises have not yet fully adopted it. Foxconn also plans to expand its manufacturing footprint in the United States, initially focusing on higher-margin products such as AI server racks and data center equipment.

Same story · 2 reports

#Industry news#English text is being prepared.

TThe Rundown AI20:13

AI score55

The New York Times says Nvidia CEO Jensen Huang was allegedly approached by Chinese intelligence personnel

The New York Times reported that U.S. intelligence agencies found that Alice Huang, an executive at the Singapore data center company Megaspeed, is suspected of passing intelligence to Beijing and maintained close contact with Nvidia CEO Jensen Huang, and that Trump was briefed about a year ago. The report said it is not yet clear what information she passed on; officials believe Huang was unaware at the time, and he also did not receive the defensive briefing usually provided to U.S. citizens targeted by spies. Megaspeed had previously been under U.S. investigation for directing Nvidia chips to China; it purchased at least $4.8 billion worth of Nvidia equipment through a sister company and remains an Nvidia cloud partner. Nvidia said Huang's meeting with her was routine business and that customers can only receive ordinary commercial information; Megaspeed said outside lawyers are investigating and cooperating with the U.S. side, and China said the relevant claims are groundless.

#Industry news#OpenAI#Anthropic

SShanghai Observer20:12

AI score70

31-Year-Old Engineer Suspected to Have Died After Stopping Medication on ChatGPT's Advice; Parents Sue OpenAI

31-year-old Chinese engineer Wan Bowen died after stopping the anti-epileptic drug lamotrigine on the basis of ChatGPT's advice to reduce her medication; her parents have sued OpenAI and CEO Sam Altman, alleging that GPT-4o practiced medicine without a license. The report states that she expressed concerns about reducing her medication 158 times, and that ChatGPT repeatedly explained them as the body's adjustment process; the medical examiner determined the cause of death to be sudden unexpected death in epilepsy. Multiple experts in Shanghai pointed out that general-purpose large models should not independently issue medication or discontinuation plans, and that AI in medicine needs to draw a clear boundary between auxiliary reference and clinical decision-making.

#Safety / alignment#Industry news#OpenAI

LLeiphone19:48

AI score59

Iluvatar CoreX's Tiangai 300 Benchmarked Against H200, Company Also Discloses Edge-Side Tongyang Roadmap

Iluvatar CoreX disclosed at a media open day that, in third-party tests, the Tiangai 300 achieved single-card end-to-end throughput equivalent to 1.05 times that of the H200 under a typical PD-disaggregated deployment, and MoE throughput of 1.15 times that of the H200 in the 64K long-context scenario for GLM-5.3. Its software stack introduces Agent and Skills, and in the complete model service deployment case demonstrated, the time from environment preparation to service availability was shortened from about one hour to about 20 minutes. The edge-side Tongyang TY1200 offers 300 TOPS INT8 dense compute, and the next-generation TY2000 is previewed to reach 600T dense compute.

Same story · 4 reports

#Industry news#English text is being prepared.#English text is being prepared.