Russia, China and the U.S. are negotiating guardrails on AI use in nuclear command-and-control systems ahead of a trilateral summit on the sidelines of APEC, per SCMP. With disputes like Ukraine and Korea largely off the agenda, talks are focusing on AI limits and a joint declaration that nuclear war can never be won. Russia remains cautious about formal curbs.
Evercore ISI chief equity strategist Julian Emanuel said Nvidia is a core part of its AI stock basket and expects Anthropic's IPO to go smoothly.
Cisco added support for Anthropic's Claude Mythos to its AI Defense offering, helping enterprises spot AI agent vulnerabilities before attackers exploit them. Optimized for cybersecurity, the model strengthens the adaptive red-teaming engine's attack reasoning and payload generation.
Thomas Neumann argues AI agents are driving overfitting in database benchmarking, calling GenDB-V's claimed 22x-61x speedups over Umbra and DuckDB infeasible. He attributes the gains mainly to per-query auxiliary structures precomputed at load time rather than to better general query processing.
The EU has introduced revamped rules aimed at influencer marketing and deceptive online tactics, meaning platforms relying on influencer promotion or dark patterns will face stricter compliance demands.
A Google software engineer used Gemini, ChatGPT, Claude and Grok to direct 48 AI agents analyzing about 1,200 km² of NASA lidar data, spotting a previously unreported Maya site in Mexico and getting an official Mexican government reply within 72 hours.
Bloomberg reports that SAP promotes its AI tools but has not disclosed enough metrics to show whether its AI effort is paying off. The same newsletter notes IBM ended visa sponsorship for some US roles after tighter Trump administration restrictions.
GitGuardian found 31,088 secrets in Claude Code's .claude/settings.local.json files across 23,276 public GitHub repos, with 5,343 still valid, including database strings, cloud tokens and Google API keys. The “don't ask again” option stores full commands with embedded credentials; monthly leaks peaked at 12,535 in April 2026 and fell to 959 by October.
Amazon Web Services pledged $100 million over the next five years for initiatives involving AI and cloud technology in healthcare, the latest push by a major cloud provider to fund AI work in the health sector.
After months of urging US officials and Anthropic to give them access to its powerful AI models, some power companies are now using the company's cybersecurity technology to hunt for flaws in energy systems and fend off hackers.
Energy-storage startup Torus said it raised $500 million to launch a shipping container-sized product combining batteries with AI computing.
England's higher education regulator, the Office for Students, says AI has made fair assessment of coursework 'nearly impossible.' It found students using ChatGPT to write entire papers undetected without recalling basic content. Universities must return to traditional exams or face sanctions to protect degree credibility.
Bloomberg notes the AI trade is so large that bulls and bears both find support on the same day. The Philadelphia Semiconductor Index fell 12% on Thursday, nearly recovered over the next two days, then weakened again Tuesday, with news from Anthropic and CoreWeave lifting optimism.
Airbnb co-founder and Chief Strategy Officer Nathan Blecharczyk said the company is developing its own specialized AI travel agent that will eventually book both on and off its platform. In a Bloomberg interview he also discussed AI's impact on Airbnb, its return to growth and European regulation.
At the Web Summit in Lisbon, Bishop Paul Tighe presented the Vatican's stance from Pope Leo XIV's first encyclical, Magnifica Humanitas: AI development should not be left solely to a few tech companies, and a machine being able to replace a human does not mean it should.
Sinovation Ventures and 01.AI CEO Kai-Fu Lee questioned whether trillion-dollar AI valuations are justified and flagged risks of circular funding, citing cheaper Chinese and open-source models. He said businesses must reinvent themselves around AI and that Europe's biggest opportunity lies in applications, not foundation models.
A Reddit post asks why ML researchers in China can publish hundreds of papers, with students producing one a week and co-author lists reaching dozens or hundreds, concentrated in LLM and multimodal work that is heavy on experiments. The author says LLM-generated artifacts flagged with ChatGPT/Claude were still accepted at ICLR and AAAI, and asks for an insider view on publishing networks and academic nepotism.
A 13-year-old girl allegedly used Gemini on her school Chromebook to plan her suicide, but the school's monitoring systems did not detect it until it was nearly too late. Her parents' allegations expose gaps in safety monitoring and flagging for AI use in schools.
At the Community of Sant'Egidio peace meeting in Assisi, Italy, Hiroshima survivors, scientists and religious leaders debated whether humanity can control the destructive forces it creates, from nuclear weapons to AI. They noted weakening arms control, record military spending and AI's rapid entry into defense systems with unpredictable consequences.
Baidu's official account teased that Ernie 5.5 Flash will be released within two weeks, without details on specifications, capabilities or availability.
AI powered roughly a quarter of UK growth this year, according to Bloomberg economists, leaving the country vulnerable to a global downturn in technology valuations.
Sonja Laud, co-head of investments at UBS Asset Management, said the AI boom has shifted from an equity story to a debt-driven balance-sheet story, creating an interconnected web of risk that markets dangerously underestimate. She questioned whether the investments will deliver expected returns.
NVIDIA has reportedly canceled the GeForce RTX 5080 24GB launch and restored the 16GB model due to GDDR7 memory costs. The report has not been officially confirmed.
Google Arts & Culture released the first online retrospective of South African artist Dr. Esther Mahlangu for her 91st birthday, featuring over 45 curated stories, archives and murals. AI was used to recreate undocumented childhood memories from her own accounts, preserving Ndebele heritage for a global audience.
A Hacker News thread highlights that open-source projects including Immich, Meilisearch, Servo, Redox OS and stb now explicitly reject code, tests and docs produced by LLMs or tools such as Copilot and ChatGPT; some also bar AI-analyzed bug reports and PRs, responding to a surge of AI-assisted submissions.
The Motley Fool argues SpaceX, after a record $85.7 billion IPO in June, could reach a $5 trillion valuation by 2029. Q3 revenue rose 161% to $14.6 billion, with its AI unit generating $7.6 billion, overtaking connectivity; a potential Tesla merger on a ~$3.4 trillion combined value would shorten the path.
Anthropic filed with Japan's Ministry of Finance to sell 38.5-50 million shares to Japanese investors at an indicative $70-$78 each, targeting $3 billion to $3.5 billion. The Claude maker began marketing its IPO earlier this month with a fundraising cap of $3.5 billion.
IBM and Microsoft announced Imagine Enterprise, an agentic platform for building, running and governing industry-process AI agents with controls, observability and compliance support. It combines over 6,000 IBM Consulting AI agents and assets with Microsoft's AI, data, cloud and security foundation.
A panel study across fifty economies found that the diffusion of artificial intelligence is associated with lower unemployment. The result adds cross-country evidence to the debate over AI's labor-market effects, though correlation does not imply causation.
Researchers have established ethical frameworks for the use of artificial intelligence in early childhood education, aiming to set norms for a fast-growing use case and protect young children.
Load more