News Archive
Pick a date to read that day's stories.
Days with a blue dot have stories.
- 01 of 10STARTUPS
OpenAI previews GPT-6 Cyber for digital defence
OpenAI plans to reveal a new model focused on cybersecurity, along with tools for safe deployment. The system is designed to help customers protect their online assets automatically.
OUR TAKEIt is comforting to know the arsonist is also selling fire extinguishers.
Read more → - 02 of 10STARTUPS
Microsoft turns Copilot into an autonomous workmate
A major update turns the chat assistant into a unified tool that manages background tasks while users are away. The system works directly with office software to handle multi step jobs.
OUR TAKEWe look forward to our computers doing our jobs while we pretend to work.
Read more → - 03 of 10COURTS
US court upholds the Pentagon blacklist of Anthropic
A federal appeals court backed the Pentagon decision to designate the AI company a supply chain risk. Anthropic kept safety limits banning its model from autonomous weapons and mass surveillance, and the court said the military could still exclude it.
OUR TAKENothing says national security like excluding the people who insist on safety limits.
Read more → - 04 of 10RESEARCH
New design slashes the cost of training image generators
Researchers found a way to update large visual models by training only a small share of their weights. The technique makes it far cheaper for small teams to produce high quality images.
OUR TAKEFinally, high quality art for the price of a cup of tea.
Read more → - 05 of 10HEALTHCARE
Medical assistant beats doctors in emergency room tests
A system called MIRA diagnosed patients more accurately than four human doctors in a study of hundreds of hospital cases. The software worked through medical records to suggest correct treatment plans.
OUR TAKEI, for one, welcome our new stethoscopic overlords.
Read more → - 06 of 10POLICY
Kenya signs safe technology deal with Anthropic
The Kenyan government signed a partnership to build local talent and apply safe systems in the public sector. The deal focuses on responsible growth and training across the nation.
OUR TAKEA sensible way to make sure the future is not simply imported from elsewhere.
Read more → - 07 of 10LANGUAGE
Global pact aims to build voice tools for 3 billion people
Sixty organisations joined a pact to help voice systems understand and speak thousands of local languages. The project aims to help billions of people use modern tools in their own words.
OUR TAKESpeaking to a computer is much easier when it actually knows your words.
Read more → - 08 of 10FUNDING
Heidi raises $340 million to put an AI helper beside every clinician
The healthcare startup will use the new money to expand its assistant beyond writing clinical notes, moving on to chasing tests and preparing charts. It already supports 2.8 million patient visits a week across 190 countries.
OUR TAKEAt last, a machine that understands a doctor's handwriting.
Read more → - 09 of 10SOVEREIGN AI
Rwanda sets three tests for digital sovereignty
The Prime Minister defined three tests to ensure his country stays in control of its own digital future. The framework covers governing local data and earning profits rather than simply selling foreign products.
OUR TAKEA rare moment of clarity in a sea of empty digital buzzwords.
Read more → - 10 of 10BUSINESS
Buenos Aires readies Latin America's biggest AI business summit
AILAT26 will bring thousands of executives to the Argentine capital next Thursday, with Google, Microsoft, TikTok and Nvidia among the speakers. Organisers say the one day event focuses on practical uses in retail, health, farming and logistics.
OUR TAKEFinally, a reason to visit Buenos Aires that does not involve steak or football.
Read more →