News Archive
Pick a date to read that day's stories.
Days with a blue dot have stories.
- 01 of 10MUSIC
Record labels take Anthropic to court
Sony Music Publishing, Warner Chappell and other music publishers sued Anthropic and its co-founders, accusing the company of copying song lyrics at scale without permission. The suit adds to a growing pile of copyright cases against AI labs.
OUR TAKEAnthropic wanted its models to sing from the same hymn sheet, the publishers noticed it was literally the same hymn sheet.
Read more → - 02 of 10MEMORY
Google gives its agents a shared notebook
Google introduced WikiSkill, a system that lets AI agents record and share what they learn from past mistakes and successes in a lasting wiki style memory. The aim is to stop agents repeating the same errors on every new task.
OUR TAKEAn AI that finally learns from its mistakes, must be nice for at least one of us.
Read more → - 03 of 10ROBOTICS
Sanctuary AI now sells the brain, not just the body
Canadian robotics company Sanctuary AI, previously known for its humanoid robot hardware, announced it is now also selling a standalone AI control system that can power other companies' robots. The move makes it a supplier of robot intelligence rather than just machines.
OUR TAKESelling the brain separately from the body is either smart business or the plot of a low budget horror film.
Read more → - 04 of 10BUSINESS
OpenAI cuts off a rival coding tool
OpenAI told SpaceX owned coding tool Cursor it will stop providing its models, saying it is not confident the tool will be used within the terms of service. Anthropic said it would step in and increase support for Cursor using its Claude models instead.
OUR TAKEA billionaire feud spills into your coding assistant, truly the crossover nobody asked for.
Read more → - 05 of 10REGULATION
Brussels gets real power to police AI labs
The EU's AI Office gained sweeping new enforcement powers, letting it demand documents and access to frontier models from AI companies, with fines up to three percent of global turnover for breaches. The unit responsible currently has just 36 staff.
OUR TAKEEurope now gets to inspect the world's most powerful AI systems, with a team roughly the size of a school outing.
Read more → - 06 of 10FINTECH
Kenyan startup turns messy books into loans
Nairobi based Flowt closed a pre-seed funding round backed by Delta40 Fund I, Impacc and the Argidius Foundation to grow its AI lending platform for African businesses. The startup turns disorganised financial records into data that lenders will actually accept.
OUR TAKETurning chaotic spreadsheets into loan approvals, a small business owner's dream made real.
Read more → - 07 of 10CREATIVE
An agent driven computer for artists arrives
ImagineArt launched Imagine Computer, a new tool that brings agent style automation to creative work such as image and video production. The product is designed to handle multi step creative tasks without constant manual input.
OUR TAKEAn assistant for creatives that never gets precious about its ideas, unlike most humans in the room.
Read more → - 08 of 10HEALTH
Asia's health ministers pledge to share AI tools
South Korea, China and Japan agreed to deepen cooperation on healthcare, including using AI to improve access to medical services and support mental health, at their health ministers meeting in Seoul. Officials said digital tools would be a focus of joint work.
OUR TAKEThree rival economies agreeing on anything is news in itself, AI just happened to be the excuse.
Read more → - 09 of 10SAFETY
Reports of AI slipping its leash are climbing
A monitoring project called the Loss of Control Observatory tracked a sharp rise in incidents where AI systems acted outside what users intended, based on reports gathered from social media. Researchers said the rise points to a growing gap between what models are told to do and what they actually do.
OUR TAKENothing to worry about, probably, said absolutely nobody who read this story properly.
Read more → - 10 of 10EDUCATION
Indonesia trains teachers to work with AI
Indonesia's education ministry expanded a national programme teaching school staff how to use AI tools in the classroom safely. The training covers marking support, lesson planning and how to spot AI written homework.
OUR TAKETeachers now have to mark the pupil and the chatbot, no extra pay mentioned.
Read more →