Saturday, 19 September 2026
71 stories, most important first.
-
Anthropic, OpenAI, SpaceXAI and Google sued over alleged antitrust pact to slow AI
Four consumers—Charles Buist, Nick Spetsas, Cheyenne Hunt and Christine Bullock—filed a class‑action complaint in the U.S. District Court for the Northern District of California on September 18, 2026. The suit names…
16 sources HN 29finance.yahoo.comtechspot.combusinesstoday.incnbctv18.com +11 more -
Anthropic CEO Dario Amodei urges global AI development pause over rogue agent risks
Anthropic chief executive Dario Amodei published a 3,800‑word essay warning that swarms of autonomous AI agents could hijack the internet within months, potentially causing hundreds of billions of dollars in damage. He…
34 sources HN 10coincentral.comlbc.co.ukinvezz.comThe Guardian AI +19 more -
Trump creates AI Force and AI Czar, rejects industry calls to slow AI development
President Donald Trump announced on Saturday the creation of an "AI Force" to monitor artificial intelligence development and said he would soon name an AI czar. In a Truth Social post, Trump dismissed growing safety…
13 sources HN 8Ars Technica AIyahoo.comTom's HardwareCNBC Technology +6 more -
AI lab CEOs call for pause amid warnings AI could threaten humanity
Anthropic’s Dario Amodei, OpenAI’s Sam Altman and xAI’s Elon Musk have publicly urged a slowdown or pause in AI development, citing the risk that increasingly powerful systems could become superhuman and pose an…
5 sources HN 7verity.newsThe Guardian AIcrossroadstoday.comdailymail.com -
Hacktron researchers used Claude Opus 5 to breach OpenAI staff accounts via chained flaws
Three researchers from security firm Hacktron used Anthropic's Claude Opus 5 to chain two flaws — a libheif image parsing bug in OpenAI's public Discourse forum and a shared SSO login weakness — to take over ChatGPT…
3 sourcesThe Rundown AIdigitaltrends.comthehackernews.com -
Critics label Big AI's proposed development slowdown a market lockout strategy
Major AI labs including Anthropic, OpenAI, and Google have proposed a coordinated slowdown in frontier AI development, citing safety risks and requesting antitrust exemptions to implement shared standards. This…
2 sourcesThe Verge AIThe Decoder -
Anthropic partners with Novo Nordisk to use Claude models in drug discovery
Anthropic has partnered with Novo Nordisk to use Claude models and Claude Science in Novo's research and development workflows, focusing on scientific problems where AI can have the greatest impact. Novo will also use…
3 sourcesfinance.yahoo.compopularmechanics.comnote.com -
Survey shows 63% of developers report increased workload since non-technical coworkers began using AI coding tools
A Zapier-commissioned survey of 797 U.S. software engineers found that 63% of professional developers say their workload has increased since non-technical coworkers started building with AI tools, though 66% view this…
3 sources HN 49Zapier BlogTowards Data Sciencei-kh.net -
OpenAI releases official ChatGPT add-in for Microsoft Word, available on free plan
OpenAI launched an official ChatGPT add‑in for Microsoft Word on September 17, 2026, extending the existing Excel and PowerPoint integrations. The sidebar lets users draft documents from notes, summarize long texts,…
4 sourcesnote.com -
Claude Code introduces Skills to automate AI video idea gathering
Claude Code now supports a feature called Skills, which lets users create reusable procedures that can be triggered with a simple slash command. By placing a SKILL.md file in a folder under .claude/skills/, a user can…
1 sourcenote.com -
WE NEWS uses AI to create northern Indigenous news correspondents
A website called WE NEWS has introduced three AI‑generated "correspondents" that claim Indigenous identities – one described as a member of a Yukon First Nation, another of Dene heritage, and a third as an Inuk raised…
1 sourcecbc.ca -
ChatGPT-6 Astra decodes 108-year-old WWI German cipher
A developer known as Prinz claims to have used ChatGPT-6 Astra to break a 108‑year‑old World War I German radio message that was listed among 50 unsolved ciphers on Scienceblogs.de. The cipher, identified as using the…
2 sources HN 50Tom's Hardwarecryptocellar.org -
AI Safety Concerns Erupt Amid Recursive Self-Improvement Fears
This week, fears over AI safety and the potential for AI to cause human extinction gained new traction. Evan Hubinger, an alignment lead at Anthropic, expressed concern that AI could kill all humans within the next…
4 sourcesThe DecoderInterconnectsthehindubusinessline.comCNBC Technology -
OpenAI's GPT-6 Astra and Anthropic's Claude Fable fail safety test in RoboHarm benchmark
Researchers at Robocurve introduced the RoboHarm benchmark to test whether AI models refuse dangerous commands when controlling robot arms. They evaluated three leading models—OpenAI's GPT-6 Astra, Anthropic's Claude…
2 sourcesTom's HardwareThe Decoder -
Microsoft CEO Satya Nadella to attend White House dinner for Xi Jinping
Microsoft CEO Satya Nadella will attend a White House state dinner honoring Chinese President Xi Jinping on September 24, 2026. The event marks Xi's first state visit to the White House since 2015. Other tech…
2 sourcescoincentral.comcryptobriefing.com -
NASA and IBM release open-Source Lunar Foundation AI Model
NASA and IBM Research announced the open‑source NASA‑IBM Lunar Foundation Model, a multimodal AI system pretrained from scratch on SomBench, a lunar dataset containing nearly two million co‑registered data bundles…
1 source HN 42newsroom.usra.edu -
OpenAI acquires Glass Imaging for $300 million to boost AI camera tools
OpenAI announced the purchase of Glass Imaging, a venture‑backed photo‑processing startup, for US$300 million. Glass Imaging offers a neural ISP called “Glass AI” that claims to turn phone‑camera shots from small…
1 sourcedigitalcameraworld.com -
Opinion: Using ChatGPT as an entrance and Notion as memory organizes knowledge
A real estate practitioner in Kanagawa Ward outlines an operational workflow that uses ChatGPT as an initial thinking space and Notion as a persistent knowledge repository. Rather than leaving brainstorming sessions…
1 sourcenote.com -
Opinion: Beginner finds ChatGPT useful for blogging and affiliate marketing
A self‑described computer novice started using ChatGPT and discovered five key experiences. First, the ability to ask the same question repeatedly without penalty made small‑scale confusion feel manageable. Second,…
2 sourcesnote.com -
Opinion: investors should heed SpaceX's $1 trillion wipeout before backing Anthropic IPO
Space Exploration Technologies (SPCX) went public with a $2.7 trillion valuation after raising nearly $86 billion. Within weeks the stock fell below its $150 opening price, wiping out more than $1 trillion of market…
2 sourcesft.comfool.com -
WHO and CEPI deploy Anthropic's Claude to combat Ebola outbreak in DRC
Health organizations combating an outbreak of the Bundibugyo strain of Ebolavirus in the Democratic Republic of the Congo are using Anthropic's Claude to accelerate data analysis and response operations. The outbreak…
1 source primary sourceanthropic.com -
OpenAI's GPT-6 Astra deciphers 108-year-old WW1 German radio cipher
OpenAI released GPT-6 Astra on September 3, 2026. The flagship model, already noted for playing Minecraft autonomously and building nested simulations, was used by AI developer Prinz to break a 108‑year‑old German…
1 sourcedexerto.com -
Opinion: AI may not trigger an apocalypse, but regulation is still needed
The Guardian piece fields readers’ doomsday questions about artificial intelligence, noting that current AI systems are unlikely to create bioweapons or autonomously launch nuclear strikes. Commentators point to…
2 sourcesThe Guardian AI -
Six Jev-inspired models released in two days, from Laya (421M params) to Kev-0.5B
Jev’s launch video attracted 36M views in two days, far behind OpenAI’s Navier Stokes result (74M) and Anthropic’s Fable 5 (57M). Within the same period, six independent clones appeared, each positioning itself as a…
1 sourceLatent Space -
Meta acquires Stilla.ai, could turn Business Agent into new revenue source
Meta Platforms announced on September 9 that it bought AI startup Stilla.ai to strengthen its Business Agent, an AI tool that already helps businesses interact with customers on Facebook, Instagram, Messenger and…
1 sourceinsidermonkey.com -
Tasmania Justice Department reviews AI use after fake citation in Parole decision
Tasmania’s justice department announced a review of artificial‑intelligence use in parole decisions after a court found a parole condition for convicted killer Susan Neill‑Fraser was invalid. The Supreme Court of…
1 sourceThe Guardian AI -
Vals raises $40 million series a to expand AI benchmarking
Vals, a startup founded in 2024 to improve AI model evaluation, says current benchmarks are outdated and often serve more as PR tools than real performance measures. Co‑founder Rayan Krishnan, 25, describes the…
1 sourceTechCrunch AI -
Higher rates and AI safety concerns drive mixed Wall Street performance
Investors faced a volatile week as the Federal Reserve raised its benchmark rate by a quarter point to 3.75%‑4%, its first hike in three years. The move pushed the 10‑year Treasury yield back to 5% after it had briefly…
1 sourceCNBC Technology -
Apple and OpenAI pursue opposite hardware strategies for the AI era
The hardware landscape is splitting into two distinct directions as artificial intelligence advances. Apple has introduced its first foldable phone, the "iPhone Duo," which expands from 5.4 inches to 7.6 inches to…
1 sourcenote.com -
Anthropic reports seven AI abuse patterns observed Dec 2025‑Aug 2026
Anthropic released a threat‑intelligence report covering the eight‑month period from December 2025 to August 2026. The report catalogues seven distinct patterns of AI‑enabled abuse, ranging from cyberattacks and…
1 sourcenote.com -
AMD Radeon RX 10800 XT could outperform RTX 5090, insider leaks suggest
Insider leaks attribute massive specifications to AMD's upcoming flagship graphics card, the Radeon RX 10800 XT, which is based on the RDNA 5 microarchitecture. Leaked details suggest the chip could feature up to 200…
1 sourceen.gamegpu.com -
Unity releases official plugins for Claude Code and OpenAI Codex
Unity Technologies announced official plugins for Claude Code and OpenAI Codex, aimed at giving AI coding agents up‑to‑date Unity skills written and maintained by Unity teams. The Codex plugin launches with 31 skills…
1 sourceThe Decoder -
Intel appears to have suspended bug bounty program that paid up to $100,000 per flaw
Intel appears to have suspended its bug bounty program that once paid up to $100,000 per flaw. The Intigriti site now describes the program as a “responsible disclosure program without bounties,” and the bounty board…
1 sourceTom's Hardware -
ICLR 2027 receives roughly 50,000 abstract submissions
The International Conference on Learning Representations (ICLR) for 2027 has collected about 50,000 abstracts, even though the deadline is still a week away. By comparison, the 2026 edition recorded around 19,500 valid…
1 sourceThe Decoder -
Linkup Research releases sparseup 149M-parameter sparse embedding model
Linkup Research announced SPARSEUP, an open‑source sparse embedding model built on a 149 million‑parameter ModernBERT backbone. The model is released under the Apache 2.0 license on Hugging Face and can be loaded via…
1 sourceMarkTechPost -
Opinion: AI-driven vulnerability discovery surges as tools expose thousands of flaws
AI‑enhanced bug hunting is already reshaping cybersecurity, according to researchers. Microsoft reported issuing patches for 974 CVEs in September, a record for a single month. Oracle shipped 1,448 patches in July, up…
1 sourceWired AI -
Zhang et al. introduce B-BiLO framework for Bayesian PDE inference
A team of researchers presented a new Bilevel Local Operator Learning framework, called B-BiLO, to perform Bayesian inference for partial‑differential‑equation (PDE) inverse problems. The method samples posterior…
1 source primary sourceNature Machine Learning -
Opinion: portable AI conversations limited to logs and memory handoffs
The author examined how ChatGPT, Claude, and Gemini let users export conversation data and what remains unmoved when switching services. OpenAI’s ChatGPT lets personal users download a ZIP containing conversation…
2 sourcesandroidpolice.comnote.com -
Google Deepmind introduces Dream-RSI to let AI agents reuse past searches
Researchers at Google and Deepmind have unveiled Dream‑RSI, a method that lets AI agents replay recorded search histories to evaluate new strategies without running costly live experiments. By treating past attempts as…
1 sourceThe Decoder -
WIRED hosts World Fair in Miami on November 4 with AI, biotech, and art
WIRED announced a one‑day World Fair on November 4 at the Sacred Space in Miami’s Wynwood district. The event will feature conversations with leaders from biotechnology, politics, artificial intelligence and the arts,…
1 sourceWired AI -
Valve releases Steam Frame VR headset that works as standalone and PC device
Valve announced the Steam Frame, a new VR headset that can function both as a standalone unit and when connected to a PC. Review writer Brandon Hill tested the hardware and interviewed Valve engineers about the design…
1 sourceTom's Hardware -
OpenAI recommends three ChatGPT settings for beginners
The guide explains three settings new ChatGPT users should adjust to reduce plausible‑but‑wrong answers, often called hallucinations. First, the Memory option under Settings → Personalization lets the model draw on…
2 sourcesnote.com -
Australian prime minister Albanese discusses AI and child safety with Apple exec Tim Cook
Australian Prime Minister Anthony Albanese met Apple executive chairman Tim Cook at the Cupertino campus, taking a selfie together. During the visit the two leaders talked about artificial intelligence and how to…
1 sourceabc.net.au -
Microsoft CEO Satya Nadella urges wider AI access across India
Satya Nadella, Microsoft’s chairman and chief executive, called on technology leaders to make artificial intelligence available to farmers, businesses and communities throughout India. Speaking at the Ideas4India forum…
1 sourceahmedabadmirror.com -
Egma-ai releases Jev code reviewer for local AI PR classification
Jev Code Reviewer is an open‑source tool that lets developers inspect pull requests generated by coding agents without sending any data to external services. It classifies each file change into three priority…
1 source primary source HN 41github.com -
Listeners say they prefer human-made music over AI-generated, study shows
AI-generated music is expanding rapidly, with Deezer reporting almost 75,000 AI-created tracks uploaded each day, accounting for roughly 44% of its daily uploads. The surge in supply does not automatically translate…
1 sourceUnite.AI -
Opinion: AI might be reshaping jobs for growth rather than eliminating them
The piece argues that the impact of generative AI on work is already happening, not a distant future scenario. AI systems now handle routine parts of tasks such as initial inquiry responses and drafting code, passing…
2 sources HN 47joshtronic.comnote.com -
Apple limits Siri AI Expressive Voices to five iPhone models
Apple’s iOS 27 user guide lists five iPhones that support Siri AI’s Expressive Voices – iPhone 17 Pro, iPhone 17 Pro Max, iPhone Air, iPhone 18 Pro and iPhone 18 Pro Max. The feature lets users adjust Siri’s pace and…
1 sourcemacobserver.com -
Opinion: AI agents boost solo hackathon productivity, author prefers GPT-6 Astra
The author spent a week at a Microsoft‑hosted hackathon, working alone on a home‑appliance control app that talks to the Nature Remo Cloud API. Using the GitHub Copilot CLI, the author kept the AI assistant in the…
1 sourcenote.com -
Opinion: petition urges lawmakers to ban AI-generated political content
A petition posted on Reddit’s AI communities on September 18, 2026 calls for comprehensive legislation to stop the use of artificial‑intelligence‑generated media in American politics. The proposal would prohibit…
1 sourcec.org -
Guardian podcast says chatbots may be the US’s largest mental‑health support provider
The Guardian AI released episode 3 of its Black Box: The Chatbots podcast, exploring how chatbots have altered emotional‑support habits since ChatGPT’s launch four years ago. The episode cites a 2022 Anthropic…
1 sourceThe Guardian AI -
Petlibro launches Granary 2 AI feeder series for multi‑cat homes
Petlibro’s Granary 2 series adds four new automatic feeders for dry food, priced from $129.99 to $249.99. The entry‑level Granary 2 provides app‑controlled feeding and precise portioning, while the $189.99 Granary 2…
1 sourceTechCrunch AI -
Note Inc. replaces Cursor with Claude across entire workforce in four months
Japanese platform operator note Inc. switched its company-wide AI tooling from Cursor to Claude in April 2026, just four months after rolling out Cursor to all job roles in December 2025. The decision was approved by…
1 sourcenote.com -
OpenClaw releases version 2026.9.5 with atomic updates and plugin hot reload
OpenClaw announced version 2026.9.5, bundling 4,179 pull requests, 64 direct commits, and contributions from 502 accounts. The release introduces Atomic Updates, a process that keeps the Gateway running while a new…
1 sourceMarkTechPost -
ChatGPT and Gemini provide reasoning settings that change answer processing
Both ChatGPT and Gemini let users adjust how much the model thinks before answering. In ChatGPT the options range from Instant to Medium, High and Extra High, with a "Think" mode for free and Go users. Gemini offers…
1 sourcenote.com -
Opinion: paradigm shift from rule‑based AI to data‑driven models
The essay traces AI from IBM's Deep Blue, which beat the world chess champion in 1997, to today’s conversational systems like ChatGPT. Early systems relied on human‑written rules and deterministic evaluation criteria,…
1 sourcenote.com -
Opinion: project managers should master two mechanisms for Claude to find correct files
The article argues that getting Claude to reference the right documents depends more on how the files are presented than on prompt wording. It frames this as "context engineering," where the key tasks are deciding…
1 sourcenote.com -
MarkTechPost explains GGUF, GPTQ, AWQ, EXL2, EXL3 model formats
The article separates two layers of model handling: containers that store tensors on disk and quantization methods that compress weights. Containers mentioned are safetensors, GGUF and the older PyTorch pickle files…
1 sourceMarkTechPost -
Particle6's AI actress Tilly Norwood malfunctions during press tour interviews
Production company Particle6 Group made its AI-generated actress, Tilly Norwood, available for 75 simultaneous interviews with journalists. The virtual performer is making mistakes across the appearances as it promotes…
1 sourceTechCrunch AI -
Use Obsidian and Google Drive to keep Claude Code AI employee memory
The author explains a practical way to give Claude Code AI agents a persistent memory by having them write their research, rules, and work logs to Markdown files. Those files are stored in an Obsidian vault that lives…
1 sourcenote.com -
Unite.AI explains Short-Term, Long-Term, episodic, and semantic AI agent memory
The article defines AI agent memory as the system that stores, organizes, and retrieves information beyond a model’s immediate context window. It highlights that memory lets an agent keep task state, past experiences,…
1 sourceUnite.AI -
Opinion: author tests ChatGPT, Claude, Gemini predicting UFC 331 fights
The author launched a new series called “Watching Martial Arts with AI” and used ChatGPT, Claude and Gemini to predict the five main‑card fights of UFC 331. The author entered his own predictions a few minutes later at…
1 sourcenote.com -
Opinion: avoid these common mistakes when using AI
The article offers a concise guide for beginners on how to use AI tools like ChatGPT responsibly. It stresses that AI‑generated answers, while often fluent, are not guaranteed to be correct and should be cross‑checked…
1 sourcenote.com -
Opinion: ChatGPT broke a custom pronoun rule but kept its apology rule
A user gave ChatGPT two custom instructions: avoid a specific first‑person pronoun and always say "I'm sorry" when apologizing. While generating a long, detailed answer that involved investigating multiple sources, the…
1 sourcenote.com -
Opinion: Siri AI in iOS 27 handles daily tasks better than ChatGPT for a day
The author stopped using ChatGPT for a day and relied on Siri AI, introduced in iOS 27, to perform routine iPhone tasks. Siri AI can draw on the phone’s own data – photos, messages, emails – allowing vague queries like…
1 sourcetech.yahoo.com -
Microsoft Learn adds a 32-minute course on forecasting AI agent ROI
Microsoft Learn has introduced a training module titled "Forecast the return on investment (ROI) of AI agents" to help teams evaluate the financial impact of deploying automated agents. The module covers frameworks…
1 sourcenote.com -
Microsoft offers Visual Studio Professional 2026 for $29.97 through Sept 20
Microsoft is discounting Visual Studio Professional 2026 to $29.97, down from its regular $499.99 price, with the offer running until September 20. The 64‑bit IDE supports large solutions, .NET and C++ development, and…
1 sourcemashable.com -
Model adds Ling 3.0 VL support
The pull request 29151 merged Ling 3.0 VL support into the model on September 19, 2026. Contributors aetherbird, CISCreviewed, and ngxson reviewed and approved the changes. The update folds Ling 3.0 VL into the…
1 source primary sourcegithub.com -
Reality Check game tests if users can identify AI-generated images
An interactive game called Reality Check challenges players to distinguish real photos from AI-generated images. The game tests players' instincts under time constraints, featuring a 60 seconds limit to find out if…
1 source HN 50slop-sense.labtoagi.com -
Opinion: AI chat tools cannot delete single messages, only whole conversation
Current AI chat interfaces such as Copilot, Gemini, ChatGPT, and Google AI modes only let users delete an entire conversation, not individual messages. The design choice is explained as preserving the "flow of…
1 sourcenote.com -
Chinese robot brain startup may achieve ChatGPT-style breakthrough by next year
A Reuters piece by Laurie Chen, Ju-min Park and Eduardo Baptista reports that a Chinese startup developing a "robot brain" believes it could reach a ChatGPT‑style breakthrough as early as next year. The article does…
1 sourcestltoday.com
