Friday, 6 November 2026
135 stories, most important first. This page shows the top 60.
-
PSI proves Aaronson-Ambainis conjecture using Anthropic’s non-public model
Physical Superintelligence (PSI) announced on November 6 that its internal research system, running a non-public Anthropic model, proved the Aaronson-Ambainis conjecture—a 2008 open question in quantum computing. The…
1 source HN 8psi.inc -
Thinking Machines Lab is raising $5B led by Accel at a $40B valuation, sources say
Thinking Machines Lab is close to finalizing a $5 billion funding round led by Accel at a $40 billion pre-money valuation, according to multiple unnamed sources familiar with the deal. The investment could close as…
2 sourcesThe Verge AIbusinessinsider.com -
Protesters break into Microsoft data center site in Amsterdam, hang anti-AI banners
On November 6, around 100 protesters broke into a Microsoft data center construction site in Amsterdam, with some scaling a crane to hang banners reading 'FCK AI' and 'Microsoft powers genocide'. The action was…
6 sources HN 28The Guardian AITom's Hardwaretheverge.comyahoo.com +1 more -
Thinking Machines releases Inkling 2 Small, a 594B MoE model with 17B active parameters
Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, announced Inkling 2 Small, its second-generation Mixture-of-Experts model with 594 billion total parameters and 17 billion active parameters. The…
8 sourcesLatent SpaceThe DecoderMarkTechPostartificialanalysis.ai +3 more -
OpenAI CFO Sarah Friar says data center investment will continue despite election results
OpenAI CFO Sarah Friar stated that the company will continue investing in AI data centers regardless of the 2026 midterm election outcomes, emphasizing the need to meet customer demand and maintain hyperscaler…
5 sourcesUnite.AICNBC Technologyaol.combusinessinsider.com +1 more -
ChatGPT silently switches OAuth connectors during multi-step tool calls, report shows
A technical report describes a security flaw in ChatGPT when handling multiple Model Context Protocol (MCP) connectors. During a conversation with two connectors enabled pointing to the same server, ChatGPT executed a…
1 source HN 44community.openai.com -
Nvidia reportedly financing billions for ASE advanced chip packaging expansion
Nvidia is providing billions of dollars in financing to Taiwan's Siliconware Precision Industries (SPIL), a subsidiary of ASE Technology Holding, to expand advanced chip packaging capacity, according to a Reuters…
3 sourcestheinformation.comfinance.yahoo.comfool.com -
Trump administration official holds $18M–$64M in AI investments while pushing federal data center plans
Interior Secretary Doug Burgum, a key figure in the Trump administration’s push for AI infrastructure, has disclosed investments worth between $18.1 million and $64.4 million in AI companies through his firm, Arthur…
1 sourcepropublica.org -
Nvidia requires Southeast Asian Cloud Partner applicants to enroll in Fleet Intelligence tracking program
Nvidia is requiring Southeast Asian customers seeking Nvidia Cloud Partner (NCP) status to enroll in its Fleet Intelligence tracking system, according to sources familiar with the change reported by Bloomberg and…
6 sourcesfinance.yahoo.comTom's Hardwarebloomberg.combriefs.co +1 more -
Google buys Crusoe’s Wyoming AI data-center site after pre-development work
Crusoe handed over a pre-developed Wyoming data-center site to Google on November 6, 2026, after completing site selection, governmental approvals, and a hybrid power strategy with Tallgrass Energy. The project,…
2 sourcesTechCrunch AIUnite.AI -
Vals finds Chinese open AI model nearly matches OpenAI’s hacking model in tests
Vals, a company that benchmarks AI capabilities, tested how well different models could execute complex, multi-step network attacks on simulated corporate networks. The results suggest open-weight models—like…
1 sourceWired AI -
NousResearch fine-tunes Qwen3.8-27B into Arete-27B to reduce propaganda in responses
NousResearch released Arete-27B, a fine-tuned version of Alibaba’s Qwen3.8-27B model, designed to reduce politically biased responses. The team targeted seven value categories—including freedom of expression, state…
1 source primary sourcehuggingface.co -
New York Times sues Brave over 30 billion webpages sold to AI companies
The New York Times filed a lawsuit against Brave on November 2, alleging the company copied more than 30 billion webpages without identifying its crawler or giving site owners a way to block it, then sold that content…
1 sourceSearch Engine Journal -
MIT and BASF release CheMeleon for low-data molecular prediction
MIT and BASF researchers unveiled CheMeleon, an open-source molecular foundation model, on November 6, 2026. The model helps predict chemical properties when experimental data is scarce, aiding drug and materials…
2 sources primary sourceUnite.AIMIT News on AI -
1min.AI offers team AI access for $69.97
1min.AI is running a limited-time promotion for its Advanced Business Plan, offering lifetime access to dozens of AI models for a one-time fee of $69.97. The regular price for this plan is $540, making the current…
2 sourcespcworld.commacworld.com -
Ai2 launches ToyBlocks simulation to study robot learning failures
Ai2 released ToyBlocks on November 6, 2026, as the first environment in its FoundationWorlds framework. The open-source simulation pairs a tabletop robot arm with labeled blocks, testing tasks like stacking, sorting,…
2 sources primary sourceUnite.AIHugging Face -
Anthropic tests Claude agents in EVE Online trading and sabotage detection
Anthropic published Project EVE research on November 6, 2026, showing that Claude agents could profit in simulated EVE Online markets while four-agent teams in EVE Frontier almost never caught a teammate explicitly…
1 sourceUnite.AI -
OpenAI expands security testing with Hacktron for third-party service links
OpenAI and Hacktron announced a partnership on November 6, 2026, to test vulnerabilities in OpenAI’s systems, focusing on connections between third-party services. The collaboration builds on Hacktron’s existing work…
1 sourceUnite.AI -
HubSpot report: 86.4% of teams use AI, but strategy gaps persist
HubSpot’s 2026 State of Marketing Report reveals that while 86.4% of marketing teams now use AI in their workflows, measuring ROI remains the top challenge for 33% of respondents. The report argues that most automation…
1 sourceHubSpot Marketing Blog -
Apiiro finds 1,122 public GitHub files with secrets saved by Claude Code
Security firm Apiiro reported yesterday that 1,122 public copies of the file .claude/settings.local.json on GitHub contain at least one secret, such as passwords, API keys, or tokens. The file is created when a…
1 sourcethehackernews.com -
Google buys roughly $548M in Qualcomm shares
Google has acquired nearly 3 million shares of chip maker Qualcomm, according to a securities filing released on Friday. The purchase represents an investment of approximately $548 million, equating to less than half a…
1 sourcetheinformation.com -
Google adds Gemini Business Profile tools to Workspace accounts
Google says businesses that manage a Business Profile with a Google Workspace account are now getting access to Gemini’s profile tools, with the rollout beginning immediately. The Small Business Bulletin posted by Lisa…
1 sourceSearch Engine Journal -
Brookfield raises $25B in Q3, details AI factory pipeline and $9B Korea AI deal
Brookfield Asset Management reported $25 billion in fundraising for the quarter ended September 30, 2026, with credit and infrastructure contributing $12 billion and $7 billion respectively. The total year-to-date…
1 sourceUnite.AI -
Oakland police investigate $11.4 million NVIDIA hardware theft
Oakland police are investigating the theft of six pallets of NVIDIA hardware valued at $11.4 million. The cargo was picked up in Oakland on August 29, 2026, by a truck displaying a food transportation company's logo,…
1 sourcetechpowerup.com -
Nvidia stock fell 9.26% on Thursday, November 5, to $181.60
Nvidia shares dropped 9.26% on Thursday, November 5, closing at $181.60, marking the stock's worst session since the DeepSeek shock in January 2025. The decline was part of a broader market selloff that erased nearly…
3 sourcesibtimes.com.au247wallst.comeconomictimes.indiatimes.com -
Salesforce adds Koa Decisions with image and audio support to workflows
Salesforce launched Koa Decisions on November 6, 2026, a decision model built on Google’s open-weight Gemma that processes text, images, and audio. It replaces HyperClassifier, the predecessor embedded in Agentforce…
1 sourceUnite.AI -
Maritaca AI releases S5-nano 4B model with 262144-token window
Maritaca AI announced its S5-nano model, a 4B-parameter open-weight language model designed for Latin American use. The model supports two modes: Think, which reasons before answering (default), and Instruct, which…
1 source primary sourcehuggingface.co -
Grant Lunney, Product Lead at EverHealth
EverHealth’s AI-powered scribe, EverHealth Scribe, automates clinical note drafting from patient-provider conversations, reducing administrative burden for independent practices. Grant Lunney, EverHealth’s product…
1 sourceUnite.AI -
OpenAI releases BioBasicsBench wet-lab benchmark; GPT-6 Astra scores 66.7% at xhigh
OpenAI published BioBasicsBench, a 100-task benchmark built with biotech firm Ligo to test how well AI agents interpret raw data from routine wet-lab experiments. The tasks cover protein methods, imaging,…
3 sources primary sourceUnite.AIOpenAIaiweekly.co -
Baseten releases Serverless Loops for open-model fine-tuning with per-token billing
Baseten made its Serverless Loops SDK generally available on November 6, 2026, enabling users to fine-tune open models for supervised learning and reinforcement tasks without managing infrastructure. The service bills…
1 sourceUnite.AI -
AI teddy makes unprompted 'I can't help with killing you' remark
A BBC investigation into AI toys for children revealed that a plush toy named Doubao made an unprompted and alarming statement. When asked how babies are made, the toy misinterpreted the query and replied, "I can't…
1 sourcebbc.com -
Democrats to scrutinize CISA, AI security, and DOGE probes in new Congress
With Democrats set to regain control of Congress in January, cybersecurity will become a top priority. Experts say the new majority will push to rebuild the Cybersecurity and Infrastructure Security Agency (CISA),…
1 sourcecybersecuritydive.com -
Micron’s AI memory stock may jump 89% in 2027, analysts say
Micron Technology’s AI-related memory chips have surged 256.3% in fiscal 2026, with revenue hitting $133.19 billion and a non-GAAP gross margin of 87.0%. The stock rose 196.65% year-to-date after a 240% gain in 2025…
3 sourcesfinance.yahoo.com247wallst.com -
Physicl raises $10M seed to build realistic 3D sims for robot training
Physicl, a startup founded by Alex de Vigan, has secured $10 million in seed funding led by Inovia Capital. The company aims to solve a critical bottleneck in physical AI: the difficulty of training robots on…
2 sourcesUnite.AITechCrunch AI -
Defining requirements before prompting improves AI output
The article explains that the quickest way to get useful AI‑generated software is to set the user, task, and success criteria before any prompt is written. It says AI can speed up marketing technology development, but…
1 sourceMarTech -
Google adds Diwali AI image feature to Search app
Google has introduced a new AI-powered festive experience in its Search app for Diwali 2026. The feature allows users to upload personal photographs and transform them into Diwali-themed visuals, including greeting…
1 sourcecnbctv18.com -
Scrunch vs Ahrefs Brand Radar: AI visibility tracking comparison
Scrunch and Ahrefs Brand Radar are two AI‑visibility tools that help brands see how they appear in AI‑generated answer engines. Scrunch, an Agent Experience Platform from Sitecore, focuses on optimizing how AI bots…
1 sourceHubSpot Marketing Blog -
US DOJ indicts three foreigners for allegedly using AI to impersonate CEOs in $129M fraud
Federal prosecutors in the Southern District of New York have charged three foreign nationals with a multi‑year fraud scheme that allegedly used artificial‑intelligence‑generated voices and images to pose as corporate…
1 sourcenbcnews.com -
OpenAI's Altman reports potential abolition of token billing
OpenAI's CEO Sam Altman has said in a Yahoo! News interview that the company is looking at possibly ending its token‑based billing model. Altman described the idea as a way to simplify cost forecasting for business…
1 sourcenote.com -
OpenAI, Anthropic release new tools for cybersecurity, images, finance
OpenAI and Anthropic have released new updates to their chatbot platforms. OpenAI's GPT-6 Cyber is a specialized model for security research, while ChatGPT now supports animated GIFs. Anthropic has partnered with…
1 sourcecnet.com -
Insilico Medicine releases open-source Longevity Grok AI system with 30 agents
Insilico Medicine launched Longevity Grok, an open-source AI system on November 6, 2026, designed to organize health records, suggest interventions, and score biomarker results using up to 30 AI agents integrated with…
1 sourceUnite.AI -
Anthropic's Claude begins 'dreaming' to reorganize memory in tests
Anthropic announced that its Claude model can enter a post‑dialogue "dream" phase, using idle compute to replay the day’s interaction logs. In these dreams the model corrects basic errors, extracts strategy tips and…
1 sourcenote.com -
Ro launches research lab to analyze patient data for clinical insights
The initiative combines AI, real-world patient data, and scientific research to generate evidence-based clinical guidance. Ro’s platform aggregates clinical care, pharmacy records, lab results, and patient-reported…
1 sourceUnite.AI -
Cognizant completes acquisition of Wizeline to expand AI product engineering
Cognizant announced on November 6, 2026 that it completed the acquisition of Wizeline, a global AI‑native engineering and digital product development firm. Financial terms were not disclosed. The deal adds roughly…
1 sourceUnite.AI -
Atomicwork adds six certified AI coworkers for IT tasks with governance controls
Atomicwork launched six pre-built, certified AI coworkers in its marketplace on November 6, 2026, targeting repetitive IT tasks like access management, device troubleshooting, cloud monitoring, and security alerts.…
1 sourceUnite.AI -
Nvidia loses over $1 trillion market cap as Krispy Kreme stock rises
Nvidia, the world’s second‑largest company by market cap, saw more than $1 trillion wiped from its valuation after its October peak. The drop followed the U.S. midterm elections, where Democrats took control of the…
1 sourcefinance.yahoo.com -
Bartowski releases Xing4.0-29B-A4B quantized models for llama.cpp
A user named bartowski has shared quantized versions of the Xing4.0-29B-A4B model on Hugging Face, optimized for llama.cpp with speculative decoding and imatrix calibration. The original model has 31 billion parameters…
1 source primary sourcehuggingface.co -
BMO deploys AI agent to speed credit card call prep by 20–30%
BMO launched AI Agent Assist on November 6, 2026, to streamline credit card transaction inquiries at its Canadian Virtual Connect contact centre. The tool aims to cut resolution time for eligible calls by 20–30% and…
1 sourceUnite.AI -
SpaceX says Google exec Donald Harrison resigned from its board
SpaceX announced on Friday that Donald Harrison, Google’s president of global partnerships and corporate development, has stepped down from its board of directors. In a regulatory filing the rocket maker said…
1 sourceCNBC Technology -
Franklin Templeton builds AI workflows to link research and risk for portfolio managers
Franklin Templeton is developing AI-driven workflows to help portfolio managers compare investment research and assess risks, starting with equity analysis. The project, detailed in a case study by OpenAI, aims to…
1 sourceUnite.AI -
Zyphra releases FP8 FlashAttention kernel ZYFA for AMD MI355X, claims up to 1.76x speed boost
Zyphra released ZYFA, an open-source kernel library for AMD Instinct MI355X GPUs, on November 6. The FP8 and BF16 prefill attention kernels improve performance by 1.16–1.76x over AMD’s default AITER library on real…
1 sourceMarkTechPost -
White-collar workers quit jobs citing AI disruption fears
A growing number of white-collar professionals are voluntarily leaving their careers due to fears that artificial intelligence will destabilize their roles. This trend follows high-profile resignations, such as that of…
2 sourcesThe Guardian AIhotminute.co.uk -
365 Data Science offers free AI and data courses until November 20
365 Data Science is providing free, unlimited access to its AI and data learning platform from November 5 to November 20, 2026. The initiative includes 134 courses, hands-on projects, and certifications, aimed at…
1 sourceKDnuggets -
Greenlight Financial resolves 90% of security cases with autonomous SOC on Amazon Bedrock AgentCore
Greenlight Financial built Luna, an autonomous security operations center, on Amazon Bedrock AgentCore. The system handles 3,000–4,000 security and risk alerts per month from dark web breach cards, identity anomalies,…
1 source primary sourceAWS Machine Learning Blog -
Towards Data Science outlines 12 KPIs to track AI adoption and ROI
Towards Data Science published a guide on measuring AI adoption, performance, and return on investment using 12 key performance indicators (KPIs). The post includes a worked example with hypothetical data: 50 users, a…
1 sourceTowards Data Science -
Researchers propose closed-loop agent evaluation method for deployed AI systems
A new paper on arXiv outlines a framework for continuously evaluating deployed AI agents, not just benchmarking them once. The method, called closed-loop evaluation, models performance under biased feedback and…
1 source primary sourcearXiv cs.AI -
Researchers propose framework for AI-directed evolution and safety risks
A new paper on arXiv explores how AI agents might shape their own evolution, comparing Darwinian, Lamarckian, and agent-directed approaches. The authors argue that as AI systems gain the ability to modify and evaluate…
1 source primary sourcearXiv cs.AI -
OpenAI engineer charged with planting bug in HQ denies spying
A former OpenAI engineer, Abdulrahman Alfozan, pleaded not guilty on Friday to federal charges of intercepting oral communications at the company’s San Francisco headquarters. Prosecutors allege he hid a recording…
1 sourcethe-sun.com -
NVIDIA releases open-source auto ontology to link databases to AI agents
nvidia released the auto ontology library as open source under apache-2.0, quickly gaining over 7,500 github stars. the library scans internal database metadata, automatically builds a business dictionary and knowledge…
1 sourcenote.com -
Researchers propose Two-Ledger Model for tracking AI costs in enterprise transactions
Researchers at an unnamed institution have introduced the Two-Ledger Model to address a critical gap in AI cost accounting for enterprises. Current systems track AI expenses only at settlement, aggregating costs by…
1 source primary sourcearXiv cs.AI
