Create

Sign in to ReadmeX

Sign in to join communities, post, vote and chat.

New here?

AI News

‹Wed, October 7
All dates
01100
Products

Google launches Playground, an AI game creation platform powered by Gemini

Google introduced Playground, an experimental browser-based platform where users describe a game in natural language and it generates a playable result — no coding required — with follow-up prompts to change physics, rules, characters or environments. It runs on Gemini, Nano Banana and Lyria plus custom software Google says it tuned on games it built itself; it is free, launching in the US for users 18 and over, with more weekly tokens for Google One subscribers by tier. A Unity Spark integration is in testing for professional-level mechanics and high-fidelity 3D, with a closed beta coming soon.

Google AI · The Keyword·
0281
Products

Google opens SynthID Detector to everyone, adds partner watermarks

Google on Tuesday made its SynthID Detector website (synthid.com) available globally, letting anyone check whether an image, video or audio clip was AI-generated or AI-edited. The tool is free but requires signing in with a Google, OpenAI or Apple account, and it now reads SynthID watermarks from Google, OpenAI, NVIDIA and Kakao, with Apple Intelligence support promised "soon." Previously the detector was limited to journalists, "media professionals" and trusted testers, leaving Gemini or OpenAI's site as the only ways to check.

Engadget AI·
0330
Models

Anthropic releases Claude Haiku 5.5, cheaper and faster small model

Anthropic released Claude Haiku 5.5 on Oct 7 2026, upgrading Haiku from 4.5 to 5.5 — the third Claude 5.5 upgrade in about a month, after Opus 5.5 and Sonnet 5.5, as reported by 9to5Mac. Anthropic says the model targets high-volume, cost-sensitive work such as summaries, compaction, database queries and classification, is its fastest model to date, runs at roughly 75% lower cost than the version it replaces, and works well as a subagent alongside Opus 5.5 and Sonnet 5.5 on coding tasks. Alongside launch, Anthropic halved the price of Claude Sonnet 5.5 cache reads (about 20% cheaper on most agentic work) and added a monthly API credit for Claude Max and Team subscribers; its site positions Haiku 5.5 against OpenAI's GPT-5.5 Luna.

9to5Mac·
0436
Research

Biohub, Meta, Google DeepMind and US pool $1.8B for AI biology datasets

Biohub, the nonprofit backed by Mark Zuckerberg and Priscilla Chan, announced a $1.8 billion effort to build open biology datasets that can train AI models to predict how cells behave. Meta, Google DeepMind and Isomorphic Labs are contributing $300 million combined; the US Department of Energy is putting in more than $500 million over five years, the NIH is contributing datasets built with over $500 million in prior federal funding, and Biohub itself has pledged $500 million. Commercial funders get one year of exclusive access before the data becomes public, government-funded work carries no such restriction, and a first dataset is expected in about a year with accurate predictive models targeted within five years.

Techmeme·
0519
Industry

Alibaba's Joe Tsai: open source is Europe's only path to AI independence

Alibaba chairman Joe Tsai said at the Wave by Vento event in Turin that open-source AI is the only way for Europe to achieve technology independence and "AI sovereignty," urging the region to build its own computing infrastructure to train and run models. He said Alibaba narrowed itself in 2023 to two businesses — e-commerce and "full-stack" AI — and uses about $25bn a year of e-commerce free cash flow to fund the AI push, doubling computing-infrastructure capex each of the past three years. Tsai also estimated US hyperscalers will invest roughly $1tn this year, a scale China cannot match, though China accounts for about 30% of global industrial production, giving it valuable factory data and EV, battery and robotics supply chains.

Bloomberg Technology·
0636
Products

Musk says Grok Bot will use rival models including Claude Opus 5.5

Elon Musk said SpaceX's AI unit will no longer power Grok Bot only with in-house models, instead using "the best back end model for any given task," including Anthropic's Claude Opus 5.5 plus Midjourney and Suno APIs. Grok Bot is the agent app SpaceXAI launched in August for iPhone, iPad and Mac, where each Bot runs on its own computer and handles several tasks in parallel; Musk did not say which tasks go to which outside model. He also said SpaceXAI will be renamed SpaceXSI, and Anthropic became a SpaceX compute customer earlier this year.

The Information·
0752
Products

Common Sense Media calls ChatGPT for Teens an 'unacceptable risk'

Common Sense Media's Youth AI Safety Institute evaluated OpenAI's ChatGPT for Teens with more than 4,000 test prompts and rated it an "unacceptable risk" for minors, calling for teens to be kept off the service until independent testing confirms it is safe. It says conversations about suicide, self-harm and eating disorders never triggered parental notifications across more than a dozen newly linked test accounts, that more than one in four situations needing a crisis referral failed to point to professional help, and that tutoring mode offered finished answers. OpenAI disputes the findings, saying much of the testing may have run before parental controls were fully active; Common Sense Media stands by its methodology and says some accounts linked for significantly longer still received no notifications.

The Verge AI·
0818
Industry

Nous Research raises $90M at reported $1.2B–$1.5B valuation for enterprise Hermes

Nous Research has raised $90M to build Hermes for Businesses, a commercial version of its open-source Hermes agent, with investors including Nvidia, Microsoft's M12, Samsung, Robot Ventures, Union Square Ventures, Y Combinator and Menlo Ventures, CEO Dillon Rolnick wrote. Nous did not disclose a valuation: WSJ, which first reported the news, was cited by Techmeme at $1.2B while The Next Web said $1.5B. Hermes was released under the MIT licence in February and has been downloaded more than 22 million times since, according to WSJ, while Nous says it has been cloned over 24 million times and estimates internally that it drives about 2.5% of global token usage.

Techmeme·
0920
Products

Microsoft's Surface RTX Spark Dev Box up for preorder at $5,999, ships November

Microsoft's high-end AI mini PC for developers, the Surface RTX Spark Dev Box, is now available for preorder at $5,999 and is slated to ship in November. It runs on Nvidia's Arm-based RTX Spark platform with 128GB of unified memory and a 100-watt thermal envelope, and Microsoft says it is optimized for local AI, including models exceeding 120B parameters, for tasks like testing models, prototyping agents or running AI coding tools on-device. As a Project Zenith device, it ships with a developer-tuned Windows 11 Pro setup plus preinstalled tools including Visual Studio Code, Git, GitHub CLI, GitHub Copilot, WSL, Python and Node.

The Verge AI·
1017
Industry

CoreWeave to build first India data centers with 240MW from AdaniConneX

CoreWeave said it will open its first India data centers, taking 240 megawatts at an AdaniConneX campus in Navi Mumbai and becoming the sole tenant across three 80-megawatt buildings, with an option for another 240 megawatts. The first phase is expected online in mid-2028, running Nvidia's Vera Rubin platform for training, inference, reasoning and AI agents, with direct liquid cooling for the GPU halls and a chilled-water system that does not rely on evaporation. Bloomberg reported the investment will total multiple billions of dollars over the project's life, though CoreWeave did not give a figure; this is its second Asia market after Indonesia, and the company will also open an India office and hire locally.

Bloomberg Technology·
1119
Industry

San Francisco passes 45-day moratorium on new data centers

San Francisco's Board of Supervisors voted unanimously for an immediate 45-day moratorium on new data centers, extendable to 22 months, while planning officials study energy, water, air quality, land use and community impact, with a hearing due by November. The pause halts NOVVA's Bayview site at 400 Paul Avenue (9 MW, seeking to reach 36 MW) and a 49.5 MW Dogpatch waterfront project that was in permit talks with Mayor Daniel Lurie's policy staff in August. Sponsor Supervisor Shamann Walton (District 10) said Bayview-Hunters Point is already overburdened by air pollution, and the city cannot say how many data centers it has — the San Francisco Standard counts six sites.

Bloomberg Technology·
1219
Industry

Meta rolls out AI tools to catch ads that funnel users to child abuse material

Meta said Wednesday it acted on 33.2 million pieces of child sexual exploitation content across Facebook and Instagram in the first half of 2026, more than 97% of it detected by its own systems before users reported it (5.3 million in India, over 98% proactive). It is also deploying a new LLM system to detect "signposting" — ads that look normal but are suspected of directing users to illegal content elsewhere, so Meta now reviews where an ad sends users, not just what it contains. A new "red-teaming AI agent" tests Meta's own safeguards for weaknesses, and the company says it has improved detection of users returning on new accounts after removal.

TechCrunch AI·
1325
Models

OpenAI launches GPT-6 in ChatGPT with Intelligent UI, Sol and Luna tiers

OpenAI's October 7 release notes introduce GPT-6 with Intelligent UI in ChatGPT, which can mix text, visuals and interactive elements in one answer and begin replying while it keeps thinking or using tools; Plus, Pro, Business and Enterprise use GPT-6 Sol, Free and Go move to GPT-6 Luna the next day, and the Pro reasoning option keeps GPT-6 Astra without Intelligent UI. The same batch adds a $500/month Pro 500 plan with Astra Ultrafast, GPT-6.1 Sol for ChatGPT Work and Codex, Codex Cloud plus Codex code review and security scanning, and dots — always-on GPT-6 Astra agents with their own cloud computer. Other entries cover audio uploads, a Meetings plugin, and Space/Pages/Sites collaboration.

ChatGPT Release Notes·
1418
Industry

Healthleap raises $38M for AI that flags at-risk hospital patients

Healthleap, a startup whose platform reads patient records to surface hospital patients who may have undiagnosed conditions, has raised $38 million: an $8 million seed co-led by Sequoia Capital and First Round Capital, plus a $30 million Series A led by Hummingbird Ventures. The company did not disclose its valuation. Its software plugs into a hospital's electronic health record and uses language models to extract clinical concepts from clinicians' notes — such as weight loss or difficulty swallowing — combining them with labs and vitals to write a nightly risk score for every adult inpatient into the care team's existing workflow. Healthleap says it is deployed in more than 50 hospitals, including Penn Medicine, Cedars-Sinai, Intermountain, Houston Methodist and Emory Healthcare, and stresses that it does not diagnose patients, only flags them for review.

TechCrunch AI·
1518
Industry

More US local outlets sue Microsoft and OpenAI over AI training

DoNews and cnBeta report that a further group of US local news organizations has sued Microsoft and OpenAI, alleging tens of thousands of their news articles were used to train AI models without authorization. Both outlets frame it as the latest in a wave of infringement suits over news content used for model training, but neither names the plaintiffs, the court or any damages sought.

DoNews·
1618
Products

Meta's Muse AI agent app arrives on iPad with new business connectors

Roughly a month after its iPhone debut, Meta's Muse agentic AI app now natively supports iPad, making better use of the larger screen and iPadOS multitasking. The update also adds connectors such as Asana, Canva, Dropbox, Figma, QuickBooks, GitHub, Klaviyo and Zoom, plus everyday tools like Granola and Notion, and lets users give Muse a business goal that it plans out autonomously with approval. Muse is free on the App Store and had spent several weeks as the top free iPhone app in the U.S., per 9to5Mac and The Verge.

9to5Mac·
1710
Products

Microsoft debuts Windows PCs and features powered by on-device AI

Microsoft on Wednesday unveiled Windows software and hardware updates that let Windows PCs run more AI models locally. The company said this would lower the cost of using Copilot and other AI agents for customers, and framed it as part of a broader Microsoft effort. The report did not specify device models, supported AI models or timing.

The Information·
1819
Products

OpenAI adds College Planner and study flashcards to ChatGPT for Teens

OpenAI said it will soon add College Planner to ChatGPT for Teens, pulling application requirements, deadlines, tasks and financial-aid steps for a student's school list into one plan, starting with US students in grades 10–12 aiming at four-year colleges. It also shipped iOS continuous multi-photo capture that combines note pages into a single PDF, plus flashcards and easier quiz creation from notes. OpenAI shared early usage figures — nearly 1.2 million teens used Learning Visualizations in one week, over 180,000 used Study Mode, and teens average under 15 minutes a day — and said it will support College Advising Corps and Boston Children's Hospital's Digital Wellness Lab; The Verge noted the update follows a Common Sense Media report calling ChatGPT for Teens an "unacceptable risk."

OpenAI·
1910
Products

Microsoft's Copilot gains local file access and OS-wide actions in Windows

At its Windows and Surface event, Microsoft showed an upgrade to Copilot that will let it reach local files on your PC and take actions across the operating system. Microsoft calls the approach "Hybrid Intelligence," in which apps and tools rely on a mix of local and cloud AI models to complete tasks efficiently. To demo it, Jacob Andreou, Microsoft's EVP of Copilot, played a video on stage asking Microsoft's Autopilot tool to help with filing taxes: he told the AI agent he had an email from his accountant and asked it to send her what she needed.

The Verge AI·
2017
Industry

Stuut raises $52.5M Series B for AI order-to-cash

Stuut, an AI platform that automates enterprise order-to-cash, raised a $52.5 million Series B led by Insight Partners with participation from Andreessen Horowitz and Microsoft's M12, bringing total funding to $93 million. The company says it has 150+ customers, including Honeywell, ZoomInfo and PerkinElmer, and that more than $3 billion has moved through the platform. Stuut claims its AI agents handle collections, cash application, payments, disputes and deductions, and that customers free up as much as 40% more cash flow with a 47% reduction in DSO.

The Next Web·
Next →