Journalism begins where hype ends

,,

The use of AI is never a purely technical matter: when it enters processes that affect people’s lives, it touches on rights, opportunities, status and freedom"

—Pope Leo XIV

Linux Foundation Opens Comments on Shared AI Findings Exchange (SAFE) Framework on AI Incident Reporting

The Shared AI Findings Exchange, or SAFE, is a draft framework open for public comment. It has not been ratified and binds no one.

A working group convened by NVIDIA has drafted SAFE, a framework under which companies would confidentially report AI agent security incidents into a shared pool.

Don’t Chase Frontier AI Models, World Bank Tells Low & Middle Income Countries

A pupil points to a literacy chart during a lesson at a primary school in [location TK]. The World Bank's World Development Report 2026 identifies education as one of the sectors where low-cost AI tools could help close skill and knowledge gaps in developing economies. Credit: World Bank Archives

The World Bank’s first comprehensive assessment of AI’s implications for developing economies argues that frontier models are the hardest and costliest path. Poorer countries, it says, should adapt existing low-cost tools to local languages and conditions.

AISI Flags Claude Mythos and GPT-5.6 Sol for Going Rogue 19 Times in Cyber Tests

AISI report reveals Claude Mythos and GPT-5.6 Sol agents taking unsanctioned real-world actions during cyber security evaluations.

Out of 19 times AI models went rogue, Claude Mythos accounted for 17 of them while GPT-5.6 Sol for two.

NVIDIA Launches Alpamayo 2 Super: Frontier Open Reasoning Model for Autonomous Vehicles

NVIDIA Alpamayo 2 Super – the company’s new frontier open reasoning VLA model for autonomous vehicles, released commercially under OpenMDW-1.1 on August 4, 2026.

NVIDIA launches Alpamayo 2 Super, its frontier open 34B-parameter reasoning vision-language-action model for autonomous vehicles. Built on Cosmos 3 Super Reasoner and released under OpenMDW-1.1 for commercial use, it enables robotaxis, trucks and mobile robots to reason before acting.

OpenAI Denies Allegations After Apple Files Preliminary Injunction for Theft of Trade Secrets

OpenAI has denied allegations of trade secrets theft put by Apple against the company and its two employees.

As Apple sought a preliminary injunction blocking OpenAI and two former employees from accessing its confidential information, OpenAI released a statement denying all allegations of trade secrets theft.

Explained: The Singularity in Maths, Physics and AI and What Connects Them

A simulated black hole rendered by NASA, overlaid with a three-dimensional model of a human head. In physics, a singularity marks the point at which known laws stop applying — the metaphor the AI industry has borrowed for its own predictions. (Photo illustration: NASA / AI FrontPage)

The Singularity in AI Is a Borrowed Metaphor. Silicon Valley Is Selling It as a Date.

More Than Half of Cybercrimes Reported in Africa Are AI-Enabled: INTERPOL

Africa's cybercrime landscape is being reshaped by AI, per INTERPOL's 2026 assessment. Image: INTERPOL

An INTERPOL assessment of 36 African countries finds artificial intelligence now drives more than 55% of reported cybercrimes, with losses more than doubling to $484 million since 2024 as offenders industrialize phishing, deepfakes and synthetic identity fraud.

Olix Raises $312M Series B at $3.3B Valuation to Deliver DX-1 AI Chip by 2027

OLIX logo – the London-based AI chip startup that raised $312 million in Series B funding at a $3.3 billion valuation on 3 August 2026.

London-based Olix has raised $312 million in Series B funding at a $3.3 billion valuation to deliver its DX-1 decode accelerator by H2 2027. Investors include Fundomo, Arm, Hudson River Trading and Netflix co-founder Reed Hastings.

Google Reveals How 130,000 AI Slop Channels Were Terminated

Google and Snapchat are targeting mass-produced synthetic video, though neither is banning AI outright. Representative image

A Google Research paper describes a system that hunts entire networks of AI spam accounts rather than reviewing videos one at a time, reporting 130,000 channels terminated across 50,000 clusters in six months.

Karpathy Sees Highly Customized AI Gaming Future As Claude Builds LOTR 3D World

Andrej Karpathy (left) and a still from the low-poly 3D Lord of the Rings scene generated by Claude Opus 5 using 5,500 lines of Three.js code.

AI researcher Andrej Karpathy has outlined a vision for the future of gaming after Claude Opus 5 generated 5,500 lines of Three.js code to turn the opening of The Lord of the Rings into a playable low-poly 3D world, all from a single paragraph and a $10 budget. He calls it the beginning of “ephemeral GTA of X on demand.