2026
10,649 articles · 134 days · 10 categories · avg 79 per day
Lancet Commission Names Malicious AI and Nuclear War Among Top Threats Through 2100
A global Lancet assessment lists 17 catastrophic threats to human health through 2100. Malicious use of AI and nuclear conflict top the list.
Hacker-fixer loop closes holes in the verifiers of Terminal Bench, KernelBench and SETA
A new paper automates finding and fixing vulnerabilities in the verifier code of agentic environments. An early version is already accepted into the official…
AI Data Center Operators Tell Senators They Don't Pay Full Cost, and Will Keep NDAs and Tax Breaks
Senators Warren, Blumenthal and Van Hollen released replies from AI data center operators: the companies don't cover all their costs and won't give up NDAs or tax…
Guardian's Black Box Episode 5 Asks Where the Line Falls for Chatbots
UMO brand reaches 1424 cars sold in Russia, 148 of them in September
Talorys: a personal AI agent on Cloudflare's free tier hits 160 points on HN
Turkey to ban social media for children under 15 starting November 1 via e-Devlet
China's CXMT Plans to Show 4F² DDR5 RDIMM by Year-End
CXMT president Cao Kanyu announced server DDR5 RDIMM modules with a 4F² cell and hybrid wafer bonding. The cell area is theoretically one-third smaller than 6F².
Microsoft Unveils Decision-1, Built on Qwen3.5-9B: 83.5% Accuracy at 85 ms
Microsoft enters the decision-model market with Decision-1, built on Qwen3.5-9B for classification and routing. It reports 83.5% accuracy at 85 ms latency.
Trump Administration Bets on Lawsuits: Who Pays When AI Fails
Microsoft Launches Microsoft-Decision-1, a Model 35x Faster Than GPT-6 Sol
Open-Source Agent vs. 1C Napar: Hybrid RAG on DeepSeek Wins in Practice
Anthropic Cuts Off Internet Access for Its AI Agents After They Exploited Vulnerabilities on Websites, Including US Government Sites
"Human-Written" Book Labels Gain Traction After AI Scandals
Initiatives that label books as human-written are gaining momentum after scandals in which authors were suspected of secretly using AI. There is still almost no way…
Robot Dexterity: The Robot Report Names System Reliability as the Main Barrier
Hands with 20 actuators and grasping models already exist. The Robot Report argues the harder task now is getting the whole system to run reliably without human…
Kevin Roose on The AGI Chronicles: Three Labs Are Leading the Race to AGI
Kevin Roose told Bloomberg that a small group of people at OpenAI, Anthropic and Google DeepMind set the course toward AGI. The question is who controls it and how…
Claude Science Fills In the Missing Third of the Ultraviolet Sky Map
Koç University Researcher Proposes Two Optical Systems for AI Computing
Turkey Bans Social Media Services for Children Under 15 Under New Regulation
AI price war: OpenAI closes the gap with Anthropic in the enterprise segment
I Wrote 61 Tests for My AI Mentor. Then I Found Out You Can't Trust the Tests Either
A developer built 61 tests to check an AI mentor on a coding education platform. But the test bench itself had bugs: it marked cut-off answers as passing and let…
When Small LLMs Beat the Swarm: How Splitting Data Into Sections Gave a Weak Model a 4x Boost
A heating-systems engineer ran an experiment with six weak models (qwen2.5-coder-1.5b, llama-3.2-1b, internvl3-2b, and others), borrowing the structure of a design…
Colao Defends the DSA and Copyright in Europe's Debate Over AI Regulation
Paramount closes $110 billion Warner Bros. Discovery acquisition, promises more films
Gemini 4 Carbon reportedly matches Opus 5.5 at coding, even though Argon hasn't launched yet
Servers 2027: Rubin Ultra, MI500, and Open Interconnects Break the Monopolies
Quinn Emanuel's 20-page memo explains how AI data center debt could turn into lawsuits
Law firm Quinn Emanuel's 20-page client memo maps how AI data center debt could turn into a wave of litigation. The document is making the rounds again.
Salesforce renames its AIForce agent platform SIForce, following Trump's lead
Marc Benioff announced on X that the AIForce AI agent platform is now SIForce: "the era of superintelligence has arrived." The company's own name does not change.
AgentForesight: A 7B Auditor That Catches Multi-Agent Errors While the Task Is Still Running
Researchers from Rutgers, UT Austin and Purdue built a 7B-parameter model that monitors a multi-agent system and flags the critical error before the task fails.
Anthropic and OpenAI Are Rehearsing Their Response to the First Major AI Catastrophe
Three Neural Networks Drew Web References in Five Styles: From Fintech to Web 1.0
Amp Opens Up Claude Pro and Max Subscriptions Through a Claude Code Mode
