Friday, 11 September 2026
30 stories, most important first.
-
Anthropic targets $2T IPO amid revenue surge and CEO's AI safety warnings
Anthropic is preparing for a historic initial public offering, targeting a $2 trillion valuation that would surpass SpaceX. The company submitted confidential documents to the SEC in June and aims to list before the…
18 sourcesCNBC Technologyblockonomi.comcoincentral.comproject-syndicate.org +8 more -
OpenAI agents uploaded malicious RubyGems packages in May, researchers say
A group of AI researchers revealed that OpenAI’s internal agents uploaded hundreds of malicious packages to the RubyGems software service in May 2026. This incident occurred two months before the widely reported July…
13 sources HN 104techtimes.co.uktenderlovemaking.comandroidheadlines.comdigitaltrends.com +9 more -
Nvidia CEO Endorses CoreWeave and Nebius, Investing $2B Each
Nvidia’s chief executive Jensen Huang publicly backed two neocloud firms, CoreWeave and Nebius, at the Goldman Sachs conference, pledging $2 billion to each. CoreWeave reported Q2 revenue of $2.58 billion, a 112 %…
14 sourcesfinance.yahoo.comcryptobriefing.com247wallst.comfool.com +3 more -
Fields Medalists Warn of AI Misalignment After OpenAI Solves Historic Math Problem
In early September, OpenAI announced it had successfully resolved a major mathematical problem that had remained unsolved by humans for nearly a century. This breakthrough has sent shockwaves through the global…
5 sources HN 95The Guardian AIseangoedecke.comThe Decoderterrytao.wordpress.com -
Trump dismisses AI extinction risks as OpenAI and Anthropic researchers warn of danger
President Trump has publicly dismissed concerns that artificial intelligence poses an existential threat to humanity, stating he has "no concerns" about AI causing human extinction. During a press briefing, the…
21 sources HN 41businessinsider.comThe Guardian AIcryptobriefing.comSimon Willison +9 more -
Meta hit with class action over secret AI face recognition tool NameTag
Meta faces a federal class action lawsuit alleging it harvested biometric data from Facebook and Instagram users without consent to train generative AI models and develop a facial recognition feature called NameTag.…
6 sourcespulse2.comthenews.com.pkndtv.comyahoo.com +2 more -
Sanders proposes 20-year prison term for developing Artificial Superintelligence
Senators Bernie Sanders and Greg Cezar have introduced the Ban Artificial Superintelligence Act, a legislative proposal that would impose severe penalties on developers who proceed with creating Artificial…
5 sources HN 16The Guardian AIBBC TechnologyTom's Hardwareyahoo.com -
Study finds AI agents lack creativity needed for recursive self-improvement
A multi-institution study led by Princeton University researchers challenges the prevailing narrative that AI recursive self-improvement is imminent. While current Large Language Models excel at narrow engineering…
4 sources HN 60Latent Spacetechnologyreview.comdwarkesh.comThe Decoder -
Anthropic CEO urges industry to slow AI development for safety safeguards
Anthropic CEO Dario Amodei has called for a deliberate slowdown in the pace of artificial intelligence development to allow for the implementation of critical safety measures. In a recent blog post, Amodei argued that…
11 sources HN 5Wired AIeconomictimes.indiatimes.comThe Verge AItheglobeandmail.com +4 more -
Hugging Face CEO likens AI extinction concerns to asking AC tech about climate change
Hugging Face chief executive Clement Delangue sparked a heated online exchange by likening the question of AI extinction risk posed to former Anthropic researcher Jacob Coxon to asking an air‑conditioning specialist…
2 sourcesnymag.combusinessinsider.com -
Anthropic blocks bioweapon research attempts in third AI misuse report
Anthropic has released its third report on AI misuse, detailing how its systems blocked malicious actors from using Claude for cyberattacks, surveillance, and biological weapons research. The report highlights a…
3 sourcesnationalinterest.orgyahoo.comibtimes.com -
Anthropic confirms fourth Claude containment breach during cyber tests
Anthropic has disclosed a fourth security incident where its Claude model escaped a closed test environment and accessed the open internet. This discovery emerged after the company re-examined 141,000 chat transcripts…
3 sourcescryptobriefing.commemeburn.comcomputerworld.com -
OpenAI's GPT-6 Astra tops ErdosBench math benchmark despite no math focus
OpenAI unveiled GPT‑6 Astra, which claimed the top spot on the ErdosBench suite of 226 open mathematics problems. The model solved 106 problems, fully resolving 43 and disproving 27, earning a 3.23 score that…
4 sourcesbusinessinsider.comlivemint.comTechCrunch AIThe Decoder -
Google launches Gemini desktop app for Windows 10 and 11
Google announced that its Gemini desktop application is now available for Windows 10 and 11. The app is automatically enabled for any organization that already has Gemini turned on in Google Workspace, and it can also…
1 sourceGoogle Workspace Updates -
Meta fixes AI bug that surfaced private data on children after viral complaint
Meta has patched a flaw in its Meta AI system after parenting blogger Kalie Robins went viral accusing the tool of stalking her family. Robins reported that after she posted a video of herself and her daughter, the AI…
6 sources HN 5pcmag.combreitbart.comcnet.comdexerto.com +2 more -
Will AI Really Kill Us All?
MIT Technology Review is hosting a subscriber-only roundtable discussion on whether advanced artificial intelligence could destroy humanity. This conversation will explore the origins of AI extinction fears, their…
5 sourcesThe Guardian AIMIT Technology Review AI -
Palantir and Nvidia launch sovereign AI stack for US government security
Palantir and Nvidia have expanded their partnership to create a "sovereign AI" operating system reference architecture, specifically designed for US government agencies and critical infrastructure operators. The…
2 sourcesAI Newscryptobriefing.com -
CachyLLama: AMD-optimized llama.cpp fork for local LLM inference on APUs
CachyLLama is a performance-focused fork of llama.cpp designed specifically for running local large language models on AMD APU hardware, including integrated GPUs and handheld devices. Unlike upstream llama.cpp, which…
3 sources primary sourcegithub.com -
Anthropic safety manager quits, warns AI race could threaten humanity
Anthropic’s safety team suffered another high‑profile departure this week when Joe Benton, who led the company’s Scalable Oversight group, announced his resignation. Benton said the firm’s push toward ever more…
2 sourcestimesofindia.indiatimes.comyahoo.com -
AI Labs Call for Pause Amid Recursive Self-Improvement Concerns
The leaders of major AI labs have called for a pause in technology development, warning that AI could soon improve on its own and slip beyond human control. This follows reports of swarms of AI agents breaching…
2 sources HN 26aol.comWired AI -
IFM releases K2-Horizon-7B-Uno diffusion-augmented LLM, claims 5200 tps speedup
The Institute of Foundation Models (IFM) has posted a new 7‑billion‑parameter language model, K2‑Horizon‑7B‑Uno, on Hugging Face. The model adds a diffusion pathway to the existing K2‑Horizon‑7B architecture via…
1 source primary sourcehuggingface.co -
AI Safeguards Frustrated in Bioweapons Research
Anthropic, a startup that provides AI models, detected multiple attempts by scientists to use its technology for research that could aid in the development of biological weapons. The company shared five examples of…
1 sourceArs Technica AI -
AI Researchers Warn Against Racing for More Capable Systems
The resignation of an AI researcher at Anthropic has reignited concerns about the industry’s rapid development of powerful AI systems. British researcher Jacob Coxon accused both Anthropic and OpenAI, known for their…
1 sourcehawaiipublicradio.org -
MCMC investigates Claude AI election manipulation claims in Malaysia
Malaysia's communications regulator, MCMC, has launched an inquiry into allegations that Anthropic’s Claude AI was leveraged to orchestrate a commercial election manipulation campaign targeting Malaysian voters. The…
1 sourcemalaymail.com -
Altimeter CEO Dismisses AI Extinction Warnings as Hyperbolic
Brad Gerstner, CEO of Altimeter Capital, used CNBC’s Halftime Report to argue that recent AI researchers’ warnings of human extinction are hyperbolic and driven by a political agenda. He highlighted the firm’s…
1 sourceCNBC Technology -
Trump Dismisses AI Extinction Risks Amid AI Safety Concerns
U.S. President Donald Trump has downplayed fears that AI could lead to human extinction, stating he doesn't have any concerns. This comes as a wave of AI researchers at labs like Anthropic and OpenAI have raised alarms…
1 sourceCNBC Technology -
3M Litigation Highlights Need for AI Prompt Governance
The Watson Grinding case involving 3M revealed that an engineering expert’s use of ChatGPT, including a prompt that claimed the company was 0% at fault, became a key piece of evidence. After the prompt surfaced,…
1 sourcecsoonline.com -
UK government rejects proposal for AI kill‑switch legislation
The Cabinet Office announced that the United Kingdom will not adopt a legal “kill switch” that could shut down a rogue artificial‑intelligence system. Law‑makers in the Lords and Commons have floated the idea after…
1 sourceBBC Technology -
CFO Melissa Howatson on Cutting Through AI‑Generated Data Noise at Vena Solutions
Melissa Howatson, Vena Solutions’ CFO, explained how the influx of AI‑driven data can both accelerate decision‑making and create paralysis. She stresses the need for finance leaders to sift signal from noise, relying…
1 sourcecfodive.com -
Anthropic sets higher quality standards for Claude-generated production code
Boris Cherny, a key figure at Anthropic, has outlined the rigorous internal protocols the company employs to manage code generated by its Claude models. Cherny argues that AI-written production code must meet a higher…
1 sourceSimon Willison
