A for AI

News Archive

Pick a date to read that day's stories.

August 2026

Days with a blue dot have stories.

DISPATCH
Sunday, August 30, 2026
  1. 01 of 10ACCESS

    South Korea to give every citizen free AI access

    South Korea is preparing to offer its whole population free use of home grown AI tools, with no usage limits, for tasks like booking doctors and sorting tax paperwork. The scheme is part of a push to widen access to domestic AI systems rather than foreign ones.

    OUR TAKEFree AI for everyone, no small print. Somewhere a telecoms billing department is weeping quietly.

    Read more →
  2. 02 of 10SOVEREIGNTY

    India launches its own AI stack called Gnani Artha

    India's Vice President formally launched Gnani Artha, a home grown AI stack built by Bengaluru firm Gnani AI, combining a 30 billion parameter model called Evon 3.3 with a platform named Plexus. Officials framed it as a step towards secure, locally built AI for government and business use.

    OUR TAKEIndia joins the club of countries insisting on building their own AI rather than renting someone else's. Understandable, given the electricity bills involved.

    Read more →
  3. 03 of 10ROBOTICS

    Former lidar bosses raise big money for robot hands

    Singapore based Sharpa, founded by the people who built lidar company Hesai, has raised 670 million dollars to develop dexterous robotic hands. Backers include Alibaba, Tencent, JD.com and Meituan, and the hands are already being tested in real commercial settings.

    OUR TAKETurns out mapping the world with lasers was just the warm up act. Next stop, robots that can pick things up without dropping them.

    Read more →
  4. 04 of 10PARTNERSHIP

    Mistral teams up with Saudi Arabia's HUMAIN

    French AI firm Mistral has agreed a multi hundred million euro partnership with Saudi Arabia's HUMAIN, linking local data centres with model development and business rollout across the Gulf region. The deal was announced on 24 August and reported more widely this weekend.

    OUR TAKEEurope's answer to the big American labs finds a wealthy new pen pal in the desert.

    Read more →
  5. 05 of 10TOOLS

    Amazon opens up its coding agent system

    Amazon Web Services has open sourced Kiro Crew, a system for running multiple AI coding agents that work through tasks without a person checking each step. Developers can now inspect, adapt and build on the underlying code themselves.

    OUR TAKEAmazon giving away the tools for free, presumably so you will rent the servers to run them on.

    Read more →
  6. 06 of 10GOVERNMENT

    India's government wants AI to summarise Cabinet notes

    India has reopened a tender seeking AI tools to summarise Cabinet notes and run facial recognition for staff attendance. The move signals a push to bring AI into everyday government administration rather than keeping it in pilot projects.

    OUR TAKENothing says modern governance like asking a chatbot to explain what the Cabinet actually agreed to.

    Read more →
  7. 07 of 10MISINFORMATION

    Chatbots turn out to be decent at spotting propaganda

    A test by NPR found that popular AI chatbots mostly identified and debunked falsehoods spread by foreign governments, doing better than standard search results at avoiding uncritical repetition of false claims. Results varied between systems but the overall trend was encouraging.

    OUR TAKEThe chatbots outperforming search engines on truthfulness says as much about search engines as it does about chatbots.

    Read more →
  8. 08 of 10DISPUTE

    OpenAI sets a cut off date for the Cursor coding tool

    OpenAI has told SpaceX owned coding platform Cursor it will stop supplying its models from 12 November, saying it is not confident Cursor will keep using the technology within its terms of service. The move deepens the ongoing feud between Sam Altman and Elon Musk.

    OUR TAKETwo of the world's richest men, still arguing over who gets to use whose software. Truly the pinnacle of adult behaviour.

    Read more →
  9. 09 of 10LEGAL

    Music publishers sue Anthropic over song lyrics

    Sony Music Publishing, Warner Chappell and other publishers have sued Anthropic and its co-founders, alleging Claude was trained on copyrighted song lyrics without permission. The suit seeks damages of up to 150,000 dollars per work.

    OUR TAKEAnthropic built a chatbot that writes poetry, and it turns out some of that poetry had a rather familiar rhyme scheme.

    Read more →
  10. 10 of 10FUNDING

    Startup raises 43 million dollars for AI that improves itself

    San Francisco based Deep Cogito has raised a 43 million dollar Series A, led by TQ Ventures with several other backers, to develop AI models designed to refine their own performance over time. The round takes the company's total outside funding past 56 million dollars.

    OUR TAKEAn AI that improves itself sounds efficient, right up until it decides your job description also needs improving.

    Read more →