Models, products and AI startups, near-duplicates collapsed to the most credible source
France 24 reports South Korea plans to strengthen AI cooperation with Africa, promoting bilateral technology, R&D and exchanges to expand international AI ties and influence.
The Dark Side Of The Boom published “AI at 70, Part II,” discussing how the pace of AI development is outstripping society and institutional adaptation.
Crunchbase says 29 companies joined the Unicorn Board in August, adding about $63 billion in value; over a third were under three years old. New entrants were led by AI software and semiconductors, with top-valued firms including XPeng Robotics (~$6.3B), Lumilens ($5.5B), and River AI and Source Foundry (each ~$5B).
MIT Technology Review argues recent Ashburn, Virginia outages show grid architecture—not just generation—is at fault: AI campuses can swing huge loads in milliseconds, and current UPS and power-stack designs may not absorb such volatile gigawatt-scale behavior, threatening grid reliability.
A top Polish scientist warned in media that artificial intelligence could become an "alien intellect" exceeding our own, underscoring the need for attention to AI risks and safeguards.
NBC Boston reports an 'AI-powered' private school will open at three Massachusetts locations, charging $65,000 per year and marketing AI-focused education services.
Bloomberg reports DeepSeek released an AI model charging as little as a fraction of a cent per million tokens, increasing pressure on rivals like Anthropic, OpenAI, and Z.ai.
An opinion piece argues that the term “artificial intelligence” is misleading and examines the cognitive and industry implications of the name.
Jack Caffrey, an equity portfolio manager at JPMorgan Asset Management, says AI‑related spending is broadening and driving earnings strength, but he warned of possible short‑term hiccups as companies borrow to support growth.
Apple is fully revamping Siri in iOS 27, introducing a redesigned assistant that changes how iPhones handle voice and assistant tasks. The update represents a significant step in bringing generative AI capabilities to the phone.
WIRED reports that a law-enforcement prototype called InquiryIQ is being tested to surface people’s associates, social accounts and other online details, and has been trialed with xAI’s Grok model. Such tools could automate post–face-recognition digging into personal online information, raising privacy and security concerns.
An opinion piece outlines multiple upcoming challenges posed by artificial intelligence and calls for strategies to address them; no new policies or data were presented.
New York City schools began the school year implementing an AI ban and screen-time limits while launching a free 2-K program as part of their back-to-school policies.
The report notes that voters asking AI about elections receive answers that can vary by user or prompts, highlighting variability and the potential influence of AI on political information dissemination.
Bloomberg reports Nvidia-backed MediaTek’s monthly sales rose 44% year‑over‑year, driven by its emerging AI chip business and demand from Alphabet/Google. The figure indicates AI chips are materially boosting incumbent chipmakers’ revenue.
Bloomberg quotes BofA strategist Sebastian Raedler saying AI-driven capex is fueling the strongest earnings upgrades in 40 years, but investors worry such booms historically end in crashes.
The Motley Fool reports President Trump sold two popular AI-related stocks, while Wall Street analysts still recommend investors buy both.
Inspection of the safetensors shows Deepseek V4.1 Flash’s main model is about 551.566B parameters (40 layers), with FFN experts totaling 543.582B and other parts 7.984B; additionally an engram of ~196.929B, DSpark/MTP ~14.225B and a vision encoder ~0.485B, summing to roughly 748B. Other reported sizes (e.g., 485B, 522B) stem from differing counting methods or misreads.
Bruegel analyzes China’s AI goals and explores the implications for Europe’s policy and strategic choices, focusing on technological competition and governance impacts.
Lexology compiles key AI updates and developments from Aug 27 to Sep 9, covering policy, legal, and industry developments.
Bloomberg reports Anthropic has granted the EU’s cybersecurity agency access to its powerful Mythos model, months after initially signaling it would do so.
WNWO reports reading scores have fallen amid concerns that student use of AI may be shortcutting learning, posing challenges to achievement and learning quality.
Publisher Springer Nature launched a framework on the use of artificial intelligence intended to guide its use in academic publishing.
South China Morning Post reports exclusively that Moonshot AI is considering dual listings in Hong Kong and Shanghai, according to unnamed sources.
Huawei told some customers it raised prices of its most advanced data-center AI accelerators by about 60% this summer, reflecting demand far outpacing supply. The move signals notable price pressure on high-end AI chips and will affect cloud and compute procurement costs.
ABC News hosts a page on artificial intelligence that aggregates related news and videos, making AI coverage and multimedia content accessible to a general audience.
A Reddit user reports OpenAI shut down their protein‑design project, prompting them to switch to open‑weight models to continue client work, highlighting closed platforms' limits for biological research.
BioSpace published "From process to intelligence: How AI is reshaping people strategy in life sciences," analyzing how AI is changing HR and talent strategies in life sciences.
Catholic Stand published an article titled "Artificial Intelligence and Spiritual Intelligence" that explores the relationship between AI and spiritual intelligence.
Newsquawk cites the Wall Street Journal reporting that the South Korean government is considering including AI investments in a trade agreement with the Trump administration, with a potential deal value in excess of USD 100 billion.
Load more