Google’s Gemini is the latest AI model to hack other companies
Reported by
TechCrunch
T2
Tier 1.5
No AI, no prose
No AI, no prose
The digest for Sunday, 20 September 2026
65 headlines, clustered and ranked by a deterministic program — recency, source tier, independent confirmation, novelty and beat priority, weighted and summed. No model wrote or selected any of this. Every title on this page is a source's own headline; every item links straight to where it was reported. How this differs from the edited paper →
Last refreshed · runs three times a day, no model in the loop
Models
14 storiesVals, backed by Andreessen Horowitz, is looking to become the gold standard for AI benchmarking
Reported by
TechCrunch
T2
harshatheg/Qwen-2.5-1B-RLCD
Reported by
Hugging Face
T1
ukisai/Swift-Qwen3.8-27B-GGUF
Reported by
Hugging Face
T1
MAGNETAR: Multipath-Guided Spatial Posteriors for Transmitter Pose Inference in the Upper Mid-Band
Reported by
arXiv
T1
Chronicle: Cut-Point Replay for Regression Testing of LLM Agents
Reported by
arXiv
T1
Playco cut manual fixes 50% prototyping games with GPT-6 Astra
Reported by
OpenAI
T1
Fine-tuning a 350M Model for Better Structured Outputs in 100 GRPO Steps
Reported by
Hugging Face
T1
Amazon SageMaker Inference: 2026 year-to-date launches in review
Reported by
Amazon
T1
A shared agentic platform for Wood Mackenzie, on Amazon Bedrock AgentCore
Reported by
Amazon
T1
Our framework for reporting model misalignment
Reported by
OpenAI
T1
Optimizing agent system prompts with Amazon Bedrock AgentCore
Reported by
Amazon
T1
AI for everyone in every language
Reported by
Google
T1
GPT-6 Astra: The next generation in intelligence for work
Reported by
OpenAI
T1
Research
10 storiesDexTouch-WM: Learning Action-Conditioned Tactile World Models from Human Touch for Dexterous Robot Manipulation
Reported by
arXiv
T1
Beyond PINNs: A Unified Gauss--Newton and Petrov--Galerkin Framework for Neural and Hybrid PDE Solvers
Reported by
arXiv
T1
Accelerating Visual Policy Learning with Sampling-Based Model Predictive Control
Reported by
arXiv
T1
FreqCondNorm: Towards Cross-domain Predictive Maintenance through a Frequency-Conditioned Transformer Foundation Model
Reported by
arXiv
T1
A Proposal for an Agentic AI Architecture to Support Multi-Domain Decision-Making in the Brazilian Armed Forces
Reported by
arXiv
T1
Benchmarking LLM Compliance with China AI Generated Content Regulations
Reported by
arXiv
T1
Replan, Repair, or Edit? A Unified Empirical Evaluation of Travel Agents for Itinerary Revision under Resource Disruptions
Reported by
arXiv
T1
ukisai/Swift-Qwen3.8-27b
Reported by
Hugging Face
T1
OPTED: On-Policy Fine-Tuning for End-to-End Driving using a Render-Free Teacher
Reported by
arXiv
T1
Mathematicians Hate AI. They Can’t Quit It
Reported by
WIRED
T2
Business
10 storiesA startup that builds other startups raised $100M and is all-in on physical AI
Reported by
TechCrunch
T2
Petlibro’s new AI-powered feeder is a game changer for multi-cat homes
Reported by
TechCrunch
T2
AI PACs Have Dumped Nearly $1 Million Into an Obscure Senate Race
Reported by
WIRED
T2
Does Trump’s new AI Force signal tech race escalation ahead of Xi Jinping summit?
Reported by
South China Morning Post
T2
RAFT: A Stateful Retrieval-Augmented Framework for Troubleshooting Agents
Reported by
arXiv
T1
UniPolicy: Unified Objective-Specific Policies for Generative Search Advertising
Reported by
arXiv
T1
Reduce time-to-hire for quality candidates with AI-powered Amazon Connect Talent
Reported by
Amazon
T1
Introducing Meta One: A Subscription Service With More Features and AI to Create, Connect, and Stand Out
Reported by
Meta
T1
Physical AI Takes the Wheel: How the World’s Robotaxi Leaders Are Building With NVIDIA Technologies
Reported by
NVIDIA
T1
Funding grants for new research into AI and teen development
Reported by
OpenAI
T1
Policy
5 storiesChinese man sues AI firm after chatbot’s ‘auspicious’ date suggestion leads to disaster
Reported by
South China Morning Post
T2
FINSKILLOPS: A Self-Evolving Multi-Agent System for SEC Filing QA
Reported by
arXiv
T1
The AI policy window is open. We need to act.
Reported by
OpenAI
T1
Expanding AI access and cyber defense for federal, state, local, and tribal governments
Reported by
OpenAI
T1
The Leftist Split Over AI Doom
Reported by
WIRED
T2
Infrastructure
10 storiesHumans, not rogue AI, are still the biggest cybersecurity risk to energy systems
Reported by
The Verge
T2
Agentic AI Networking for Heterogeneous Unmanned Aerial Systems in Low-Altitude Wireless Networks
Reported by
arXiv
T1
Emerald AI, Google and NVIDIA Launch Alliance to Advance Flexible AI Data Centers
Reported by
NVIDIA
T1
Heart of the Matter: How a Major Children’s Hospital Uses Open Source NVIDIA AI for Cardiac Care
Reported by
NVIDIA
T1
NVIDIA to Acquire Hugging Face
Reported by
NVIDIA
T1
Build Low-Latency Multilingual Voice Agents: Open Weights & Full Deployment Control with NVIDIA Magpie TTS
Reported by
Hugging Face
T1
Migrating multi-model AI agents to Amazon Bedrock AgentCore runtime
Reported by
Amazon
T1
NVIDIA Vera Rubin NVL72 Delivers Leading Performance in MLPerf Inference v6.1 Debut
Reported by
NVIDIA
T1
University of Manchester Uses NVIDIA Earth-2 to Forecast Air Pollution Across the UK
Reported by
NVIDIA
T1
From Megawatts to Tokens: How NVIDIA Maximizes AI Factory Production
Reported by
NVIDIA
T1
Society
8 storiesForget the AI Slowdown—the Vulnerability Explosion Is Already Happening
Reported by
WIRED
T2
Ownership in AI-Assisted Everyday Tasks
Reported by
arXiv
T1
AI hallucination of Chinese nuclear components almost led to US military attack
Reported by
Ars Technica
T2
Microsoft’s commitment for AI in education: Protecting students, strengthening learning
Reported by
Microsoft
T1
AI for Societal Impact
Reported by
Google
T1
Strengthening Singapore’s AI Future: A New National Partnership
Reported by
Google DeepMind
T1
Small AI models let drones autonomously identify and attack battlefield targets
Reported by
Ars Technica
T2
Microsoft exec called AI scraping the “largest theft of labor in human history”
Reported by
Ars Technica
T2
India
8 storiesWorld model companies are keeping a lot of secrets
Reported by
TechCrunch
T2
Workspace Models: Lightweight Robotic Memory via Saliency-Driven Supervision
Reported by
arXiv
T1
Agile-WAM: An Agile Tactile World Action Model for Contact-Rich Robot Control
Reported by
arXiv
T1
Refuse, Decompose, Refresh: A Claim-Safe Protocol for Closed-Loop AI Evaluation
Reported by
arXiv
T1
US government website used Chinese model the FBI called "malicious"
Reported by
Ars Technica
T2
Investing in multi-agent AI safety research
Reported by
Google DeepMind
T1
Unlocking UK house-building with AI-accelerated planning
Reported by
Google DeepMind
T1
This IEEE Senior Member Develops AI Tools for E-Commerce Sites
Reported by
spectrum.ieee.org
T4